摘要:

424 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 ASU Highlights The ASU Review in this issue describes developments in atomic spectrometry relevant to the analysis of chemicals, iron, steel and non-ferrous metals. It was com- piled from abstracts received during the period January-December 1986 and hence follows chronologically from the equivalent review published in JAAS, Volume 1 (J. Anal. A t . Spectrom., 1986, 1, 87R). The major topics of interest reported in the review are similar to last year, although there has been undoubtedly a marked reduction in the number of con- ference and published papers on indus- trial analysis, particularly in the metallur- gical field. Developments in ICP-AES continue to be widely reported and not surprisingly, this ubiquitous technique dominated most subject areas in terms of published research. The most significant ICP-AES advances appear to be in the analysis of catalysts, micro-electronics, organic materials ( e . g . , petroleum, poly- mers and solvents) and in on-line analysis for process control. Electrothermal AAS is still the preferred technique for many trace element determinations, but ICP- MS has yet to make an impact in the general chemicals area. Unfortunately, there has been unnecessary publication of routine procedures in some areas and it was noted that many of the supposedly “new” metallurgical methods were merely modifications of established principles. David Littlejohn University of Strathclyde, UK

ISSN:0267-9477    

DOI:10.1039/JA9870200424

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 425 Conferences and Meetings New Perspectives in Atomic Spectroscopy September 3-4, 1987, Plymouth, UK A joint meeting of the Atomic Spec- troscopy Group, Western Region of the Analytical Division and the Peninsula Section of the RSC will be held to celebrate the appointment of Les Ebdon as Professor of Analytical Chemistry at Plymouth Polytechnic. The speakers will include: Professor L. Ebdon, Dr. J.-M. Mermet, Dr. R. D. Snook, Dr. J. B. Dawson, Dr. P. Jones, J. G. Williams, Dr. K. C. Thompson, H. G. M. Parry and Dr. J. M. Harnly. For further information contact Dr. N. W. Barnett, Dept. Enviromental Sciences, Plymouth Polytechnic, Drake Circus, Plymouth, Devon PL4 8AA, UK; telephone 0752 221312. 1987 Eastern Analytical Symposium September 13-18, 1987, New York, USA The 1987 Eastern Analytical Symposium will feature a special Symposium honour- ing Professor Velmer A.Fassel, Iowa State University, on the occasion of his receiving the 1987 EAS Award, This will be held on Thursday September 17th at the New York City Hilton Hotel. Participants in the Symposium include Ramon M. Barnes, Richard F. Browner, R. S. Houk, Qinhan Jin, Peter N. Keliher, Michael W. Routh, Xi-en Shen and Andrew T. Zander. Related Symposia during the meeting week include ICP-AFS, Lasers in Analy- tical Chemistry and Trace Metal Specia- tion and as previously there will be an exhibition, short courses and workshops. Further information can be obtained from Professor Peter N. Keliher, Depart- ment of Chemistry, Villanova University, Villanova, PA 19085, USA.(See page 421 for additional details of the programme.) FACSS XIV, 1987 October 4-9, 1987, Detroit, MI, USA The 1987 FACSS meeting will be held at Cob0 Hall and the Westin Hotel in Detroit, Michigan. As in the past, work- shops and short courses will be offered prior to, during and after the conference. The FACSS Employment Bureau will again be available to conference atten- dees. Centrally located at the meeting will be an exhibition of scientific instrumenta- tion, services and publications. For further information contact the publicity chairman, Dr. Stephen J. Swarin, Pub- licity Chairman, Analytical Chemistry Dept., General Motors Research Labs., Warren, MI 48090-9055, USA, telephone 313-986-0806.Analyticon 87 October 13-15, 1987, London, UK Analyticon 87 will run concurrently with the British Laboratory Week and will be held at Olympia, London. Four branches of analytical atomic spectroscopy will be covered in one ses- sion: inductively coupled plasma optical426 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 1988 Winter Conference On Plasma Spectrochemis try San Diego, California, USA January S 9 , 1988 The 1988 Winter Conference on Plasma Spectrochemistry, fifth in a series of biennial meetings sponsored by the ICP Information Newdefter , will feature developments in plasma spectrochemical analysis by inductively coupled plasma (ICP), d.c. plasma (DCP), microwave plasma (MIP) and glow and hollow-cathode discharge (GDL, HCL) sources.The meeting will convene Monday, January 4 to Saturday, January 9,1988 at the San Diego Princess resort and convention centre in San Diego. Expert short courses at introductory and advanced levels and an exhibition of spectroscopic instrumentation also will be included. __ Programme and Objectives Symposia organised and chaired by recognised experts will include the following topics: (1) Sample introduction and transport phenomena; (2) Instrumentation and automation, including on-line analysis and remote systems; (3) Excitation mechanisms and plasma characteristics; (4) Interferometry; (5) Atomic fluorescence; (6) Glow and hollow-cathode discharges; (7) Flow injection analysis; (8) Chromatography and plasma detectors; (9) Plasma source mass spectrometry; (10) Industrial applications of ICP mass spectrometry; and (11) Sample preparation and pre-concentration techniques. Six plenary and 15 invited lectures will be presented.Three afternoon poster sessions will feature applications, automation and new instrumentation. Four panel discussions will address critical development areas. Plenary, invited and submitted papers will be published as the official conference proceedings following the meeting after peer review in Journal of Analytical Atomic Spectrometry, September 1988 issue. Instrument Exhibition A three day exhibition of spectroscopic instrumentation and chemicals, electronics, glassware, publications and software supporting plasma spectroscopy will complement the scheduled sessions. Expert Short Courses Introductory and advanced four-hour short courses will be offered January 2-3 and 9,1988.Designed to provide background and intensive training in popular topics of plasma spectrochemistry , these will cover analytical applications, instrumentation, samples introduction and various techniques ( e . g . , plasma diagnostics, scientific writing, chemical and physical pre-concentration and applications of isotope dilution and tracers). Registration The conference registration fee includes a copy of the conference proceedings, abstracts, a tee-shirt and conference dinner. The pre-registration fee is $275 until October 16, 1987, after which time it will be $375. On-site registration will be $400. Discounts are provided for students, and no registration fee is required for spouses.Short-course pre-registration fee $75 for each four-hour short course, after October 16 this will be $100. Further details on all aspects of the Conference can be obtained from: Dr. Ramon M. Barnes Department of Chemistry, GRC Towers, University of Massachusetts, Amherst, MA 01003-0035, USA (413) 545-2294JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 427 emission spectrometry (ICP-OES); ICP mass spectrometry (ICP-MS); and atomic absorption spectroscopy with either flame (FAA) or electrothermal atomisation The session has been arranged by Colin Watson who will chair the meeting. ICP- OES will be discussed by Dr. E. J. Newman (BDH Chemicals). Dr. Stephen Long (AERE, Harwell) will talk about HPLC - ICP-MS. Dr. K. C. Thompson (Yorkshire Water) will offer some solu- tions to unexpected interferences in FAA.Finally the applications of elec- trothermal atomisation will be discussed by Dr. C. Fuller. Experimental Design in Chemistry Ses- sion: good experimental design enables the right data to be collected accurately and effectively and the effect of environ- mental noise to be removed or minimised. The aim of this session of Analyticon 87 is to provide an overview of the practical and theoretical aspects of good design and then to give a detailed introduction to the use of three design tools in chemical applications. These tools are: response surface methodology, simplex optimisa- tion and fractional factorial designs. The chairman of this session will be Dr. H. J. H. MacFie of the AFRC Institute of Food Research Bristol Laboratory.The opening lecture on principals of good design will be given by 0. Whelehan of the BP Research Centre in Sunbury-on- Thames. He will be followed by Dr. G. Nickless from the University of Bristol talking on response surface design. Simplex designs to optimise chemical experiments will be the subject of a lecture from Dr. K. Burton of the Poly- technic of Wales in Pontypridd. The final lecture before a discussion period will be given by D. E. Morgan, also from the Polytechnic of Wales. His lecture title will be “Fractional designs in chemistry: their use and analysis.” The session will be on the morning of Wednesday 14th October and the programme has been arranged by the UK Chemometrics Discussion Group. All enquiries should be addressed to either Beverly Humphrey or Bevan Gil- pin at Scientific Symposia Ltd., 33-35 Bowling Green Lane, London EClR ODA, UK; telephone 01-837 1212; telex 299049 G.39th Pittsburgh Conference and Exposi- tion on Analytical Chemistry and Applied Spectroscopy February 22-26, 1988, New Orleans, Louisiana, USA The 1988 conference and exhbition will be held at the Rivergate and New Orleans Convention Center. The Technical Pro- gramme will consist of approximately 30 symposia involving 900 contributed pap- ers, together with some poster sessions. For further information concerning the conference write to the Pittsburgh Con- ference, 12 Federal Drive Suite 322, Pittsburgh, PA 15235, USA. (ETA-AAS) . Fourth Biennial National Atomic Spectro- scopy Symposium June 28-July 1, 1988, York, UK The Fourth BNASS will be held at the University of York.Plenary lecturers will include Dr. A. L. Gray (University of Surrey, UK), Professor G. M. Hieftje (Indiana University, USA), Dr. N. Omenetto (Ispra, Italy), Dr. M. Thomp- son (Birkbeck College, London, UK) and Dr. A. P. Thorne (Imperial College, London, UK). Invited speakers will include Professor F. Adams (Antwerp, Belgium), Professor P. N. Keliher (Vil- lanova University, USA), Dr. S. J. Has- well (Thames Polytechnic, London, UK), Professor D. A. King (University of Liverpool, UK), Dr. D. Littlejohn (Uni- versity of Strathclyde, UK), Dr. E. J. Newman (BDH Chemicals, Dorset, UK), Professor A. Sanz-Medel (Oviedo, Spain) and Dr. B. L. Sharp (Macaulay Land Use Research Institute, Aberdeen, UK).In addition there will be submitted papers, poster sessions, an exhibition and full social programme. Further information can be obtained from the Secretary of the organising committee: Dr. R. Miller, Unilever Research, Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, Merseyside L63 3JW, UK. X CANAS and VII Polish Spectroanaly- tical Conference September 5-9, 1988, Torun, Poland The X Conference on Analytical Atomic Spectroscopy (CANAS) and VII Polish Spectroanalytical Conference, with inter- national participation, will be organised by the Commission of Analytical Atomic Spectrometry of the Committee of Analy- tical Chemistry of the Polish Academy of Sciences and by the Nicholas Copernicus University in Torun, Poland, where the Conference will be held.The following branches of spectroscopy will be covered: physical aspects of analy- tical atomic spectroscopy; optical emis- sion spectroscopy and excitation sources; atomic absorption and fluorescence spec- troscopy; X-ray spectroscopy; inorganic mass spectroscopy; electron and ion spec- trometry; instrumental neutron activation analysis; lasers in analytical atomic spec- troscopy; application of spectroscopy in trace analysis; speciation analysis; and surface analysis. All correspondence concerning the Conference should be addressed to: Dr. J. Fijalkowski, Institute of Nuclear Chemistry and Technology, ul. Dorodna 16, 03-195 Warszawa, Poland. First Rio Symposium on Furnace Atomic Absorption Spectrometry September 18-23, 1988, Rio de Janeiro, Brazil This is a preliminary announcement and call for papers for the First Rio Sympo- sium on Furnace AAS to be held at Rio Datacentro Auditorium, Pontificia Uni- versidade Catolica do Rio de Janeiro.Topics to be covered include: fundamen- tal studies; new advances in instrumenta- tion; furnace materials; sample introduc- tion, including direct solid analysis; method developments for the analysis of real samples; and novel applications. Abstracts of contributed papers (in English) should be received not later than May 15th, 1988. Further information can be obtained from Dr. Adilson Josk Curtius, Depart- amento de Quimica da PUC-RIO, Rua Marques de Sio Vincente, 225, 22453, Rio de Janeiro, RJ, Brazil; telephone (021) 2749922, extn. 574; telex (021) 31048 PUCR BR. 3rd International Colloquium on Solid Sampling with Optical Atomic Spectro- scopy October 10-12, 1988, Wetzlar, FRG The Colloquium is being organised jointly by GDCh, Fachgruppe Analytische Chemie and A.M.S.E.I., Arbeitskreis fur Mikro- und Spurenanalyse der Elemente, the organising committee consisting of M.Stoeppler, R. F. M. Herber, P. Kipper and U. Obbarius. This series of meetings is intended to focus on the state-of-the-art and progress of solid sampling with optical atomic spectroscopic methods. It will consist of invited and contributed papers and possibly posters with discus- sion time for themes of particular impor- tance. Submitted papers are being called for on: theory and instrumentation; methodology and procedures; biological and medical applications; food applica- tions; environmental applications; and product and quality control.The preferred conference language will be English, no translation will be provided. For further information contact Dr. M. Stoeppler, KFA Julich-ICH-4, Postfach 1913, D-5170, Julich, FRG. XXVI Colloquium Spectroscopium Inter- nationale July 2-9, 1989, Sofia, Bulgaria The XXIV CSI will be held at the National Palace of Culture, Sofia, Bul- garia. The scientific programme will touch upon all traditional areas of analy- tical spectroscopy. An instrument exhibi- tion is planned in conjunction with the Colloquium. For participants and accom- panying persons, there will be social events and excursions of scientific, cultu- ral and touristic interest rounding off the programme. Also planned are tours before and after the congress. The Con- ference languages will be English, French and German. For further information contact the Chairman, Associate Professor Dr. A. Petrakiev, Sofia University, Faculty of Physics, Department of Optics and Spec- troscopy, 5, A. Ivanov Blvd. BG-1126, Sofia, Bulgaria; telephone (3592) 627475.428 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 The Fourth Biermid National Atomic Spectroscopy Symposium will be held at the University of York on 1988 28 June - 1 July 1988 1988 The symposium will provide a forum where interesting and useful applications of atomic spectroscopy can be reported and discussed. In addition to plenary, invited and submitted lectures, a particular feature of the meeting will be the presentation of posters. There will also be an exhibition and a full social program for delegates and their guests. This meeting is organised by the Atomic Spectroscopy Group, Analytical Division of the Royal Society of Chemistry and the Spectroscopy Group of the Institute of Physics. Further information can be obtained from the Secretary of the organising committee: Dr R Miller, Unilever Research, Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, Merseyside L63 3JW

ISSN:0267-9477    

DOI:10.1039/JA987020425b

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

428 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 Future Issues will lnclude- The September Special Issue of JAAS will contain a selection of papers from the 1987 Winter Conference on Plasma and Laser Spectrochemistry, Lyon, France, 12th-16th January 1987. The October issue will contain the following papers. Low-pressure Discharges: Fundamental and Applicative Aspects. A Review- s. Caroli Determination of Barium, Calcium, Iron, Potassium, Magnesium and Sodium in High Purity Niobium Pentoxide by Flame Atomic Absorption Spectrometry-T. J. Chruscinska Determination of Yttrium in a Zirconia Matrix by Atomic Absorption Spectro- scopy-A. Samdi, J. Paris, J. P. Deloume and G. Duc Study of Resonance and Non-resonance Atomic Fluorescence Transitions in Plasma Spectrometry-S. Greenfield, K.F. Malcolm and M. Thomsen Mechanisms of Ionisation and Atomisa- tion of Barium in Graphite Furnace Atomic Absorption Spectrometry- Etsuro Iwamoto, Shuichi Ohkubo, Manabu Y amamoto and Takahiro Kuma- maru Measurement of the Practical Resolving Power of Monochromators in Inductively Coupled Plasma Atomic Emission Spec- trometry-Jean-Michel Mermet Sorption and Atomisation of Metallic Hydrides in a Graphite Furnace-R. E. Sturgeon, S. N. Willie, G. I. Sproule and S. S. Berman Fundamental Sample Introduction Stud- ies of a Graphite Rod Electrothermal Vaporisation Device in Inductively Coupled Plasma Atomic Emission Spectrometry-Susan M. Schmertman, Stephen E. Long and Richard F. Browner Flow Injection Ion-exchange Pre-concen- tration for the Determination of Alu- minium by Atomic Absorption Spec- trometry and Inductively Coupled Plasma Atomic Emission Spectrometry-M. R. Pereiro Garcia, M. E. Diaz Garcia and Alfredo Sanz Medel Study of a Microwave-induced Plasma (Surfatron) as a Detector in Capillary- column Gas Chromatography with Refer- ence to Pesticides-Brigitte Riviere, Jean- Michel Mermet and Daniel Deruaz Evaluation of the Grid-type Nebuliser for the Introduction of High Dissolved Salt Solutions and High Solid Content Solu- tions into the ICP-Timothy Brotherton and Joseph Caruso Atomic Spectrometry Update The Update in the October issue is- Atomisation and Excitation-Barry L. Sharp, Neil W. Barnett, John C. Burridge and Julian F. Tyson

ISSN:0267-9477    

DOI:10.1039/JA9870200428

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 429 Laser-excited Atomic Fluorescence Spectrometry in an Atomic Absorption Graphite Tube Furnace Joseph P. Dougherty, Francis R. Preli Jr. and Robert G. Michel” Department of Chemistry, University of Connecticut, Storrs, CT 06268, USA A dye laser pumped by an excimer laser was used for laser-excited atomic fluorescence spectrometry (LEAFS) in a Perkin-Elmer HGA-500 graphite tube atomiser. The tube furnace was modified for LEAFS by drilling two 4-mm holes across the centre of the tube. Laser light was either passed down the bore of the tube or across the centre of the tube and fluorescence was detected at a right angle. Atomisation was investigated with and without a L’vov platform. Resonance and non-resonance LEAFS transitions were studied.Detection limits for Ag, Co, Cu, In, Mn, Pb and TI were 8, 500,400, 80,80, 7 and 100 fg, respectively. The linear dynamic ranges (relative slope of one) for these elements extended 4-6 orders of magnitude above the detection limit. The relative standard deviation for 14 measurements of 200 pg of lead was 10%. Keywords: Laser-excited atomic fluorescence spectrometry; graphite tube atomiser; ultra-trace elemental analysis Laser-excited atomic fluorescence spectrometry (LEAFS) is one of the most sensitive techniques available for elemental analysis. The best limits of detection reported in the literature, for LEAFS, have been obtained with electrothermal atomisa- tion (ETA). Bolshov et al.,l Tilch et a1.2 and Goforth and Winefordner3 used LEAFS to probe the atom population above a graphite cup atomiser and were able to achieve femtogram detection limits for a total of 11 elements.However, Bolshov et al. 1 found severe vapour phase interfer- ences when they used the cup atomiser to determine cobalt in agricultural samples. They resorted to atomisation under vacuum to reduce gas phase reactions.4 Open atomisers (cups, rods, boats, etc.) have already been found to be impractical for atomic absorption spectrometry (AAS) due to diffusion losses5 and vapour phase interfer- ences6 that occur in the cool region above the atomisers. Modern ETA-AAS utilises enclosed graphite tube atomisers to reduce diffusion losses. Rapidly heated tube atomisers, equipped with L’vov platforms, and the use of matrix modifiers have reduced many vapour phase interferences in ETA-AAS by enabling the analyte to be vaporised at the optimised temperature under approximately isothermal con- ditions.’ Dittrich and Stark8 reported a comparison of LEAFS using a modern AAS tube atomiser (a Perkin-Elmer HGA/EA 3) with an open, rod atomiser.They concluded that the tube atomiser produced increased sensitivity over the rod. However, their LEAFS detection limit for lead was 45 times worse than that reported by Bolshov et al. 1 for LEAFS in a cup atomiser. Furthermore, Dittrich and Stark were unable to achieve linear calibration graphs ( i . e . , graphs with a relative slope of one) for LEAFS in their tube atomiser. The work presented in this paper utilised a commercial furnace atomiser (a Perkin-Elmer HGA-500) to achieve detection limits for LEAFS that were comparable to the best literature values for LEAFS in cup furnaces.The calibration graphs had a relative slope of one and extended over 4-6 orders of magnitude. The focus of this work was to apply AAS modern tube furnace technology to ETA-LEAFS. Experimental Instrumentation The instrumentation for ETA-LEAFS has been described elsewhere9 and is summarised here. An excimer laser operat- * To whom correspondence should be addressed. ing at 80 Hz with xenon chloride (308 nm) was used to pump a tunable dye laser. The dye laser output was frequency doubled or used directly depending on the atomic transition of interest. The frequency doubled output beam was approximately rectilinear in shape with a height of 4 mm and a width of 1 mm and was passed directly through the atomiser.When fre- quency doubling was unnecessary, the dye laser output was focused, with lenses, to a 4 mm diameter circle before being passed through the atomiser. The laser operating conditions for each element are summarised in Table 1. Fluorescence was measured as a function of laser power to determine if the atomic transition was saturated. An absolute saturation power can be determined from the plot of fluorescence versus laser power if the laser power is homogeneous throughout a well defined illumination volume and allowances are made for self-absorption and filter effects. For practical purposes we defined our saturation power (Table 1) in empirical terms as the laser power above which fluorescence is not linearly proportional to laser power.For the saturation studies, the laser output was attenuated either by reducing the high voltage of the excimer laser or by using a variable optical attenuator (Newport, Model 935-3, Fountain Valley, CA, USA). The two approaches to laser attenuation produced similar empirical saturation graphs for lead. Laser power was measured using a pyroelectric joule meter (Molectron , Model J3-05D W , Sunnydale , CA, USA). The detection system consisted of a monochromator, a Table 1. Laser conditions for ETA-LEAFS Saturation power or maximum SHGt Saturation power used/ Element Dye* crystal (Yes/No) pJ per pulse Ag . . , . . . DCM KDP-C No 2 CO . . . . . . R-610 KDP-C No 5 - Yes 5 Mn . . . . . . R-560 KDP-B No 3 Pb .. . . , . R-575 KDP-B Yes 2 - Yes 6 CU . . . . . . DCM KDP-C Yes 1.5 In . . . . . _ DPS T1 . . . . . . BBQ * DCM = 4-Dicyanomethylene-2-methyl-6-p-dimethylamino- styryl-4H-pyran; R = rhodamine; DPS = 4,4‘-diphenylstilbene; and BBQ = 4,4”-bisbutyloctyloxy-p-quaterphenyl. 1- SHG = second harmonic generation.430 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 photomultiplier tube (PMT), a pre-amplifier and a boxcar integrator. The monochromator was placed on its side to ensure maximum viewing of the laser illumination volume. (The slit was then parallel to the direction of propagation of the laser beam.) Data from the boxcar were collected by a computer and the transient furnace LEAFS signals were transferred from the computer to a chart recorder at a rate which allowed for the slow response time of the recorder.The temporal peak areas of the LEAFS signals were obtained using a computerised integration algorithm. The atomiser was an HGA-500 graphite tube furnace (Perkin-Elmer, Norwalk, CT, USA) equipped with an AS-40 autosampler (Perkin-Elmer). It was operated with the stan- dard internal and external argon gas flows (300 ml min-1 and 900 ml min-1, respectively) that are normally used for ETA-AAS. The internal gas flow was reduced to 10 ml min-1 during atomisation. A 5-mm hole was drilled through the graphite electrode contact and 4-mm holes were drilled through standard pyrolytically coated graphite tube furnaces (Perkin-Elmer) to accommodate the laser beam (Fig. 1). After drilling the laser holes, the graphite furnaces were pyrolytically re-coated at 2000 “C for 2 min in a 9 + 1 argon - methane atmosphere.Furnace tubes could be used for 500 heating cycles without loss of sensitivity. Each heating cycle included drying of the sample at ca. 200 “C, atomisation at the optimum temperature (see Table 2), and a “clean-out” step at 2600 “C. Solid pyrolytic graphite platforms (Perkin-Elmer) were reduced in height to prevent stray laser radiation from being reflected into the detection system. This was particularly important for resonance LEAFS. Standard Solutions Standard solutions were made daily by serial dilution of 1 mg ml-1 stock solutions in 0.04 M Ultrex grade nitric acid (J. T. Baker Chemicals, Phillipsburg, NJ, USA) in a class 100 clean-air environment, The stock solutions were made from the high-purity metal or metal salt (SPEX Industries, Metu- chen, NJ, USA).For each atomisation, a sample aliquot of 20 p1 was delivered by the AS-40 autosampler or manually with an Eppendorf 4700 pipette (Brinkmann Instruments, West- bury, NY, USA). Detection Limits and Linear Dynamic Ranges Detection limits were based on three standard deviations of 16 measurements of the blank. The measurements of the blank were performed either with the laser tuned to the analytical wavelength (the “on-line” measurement) or with the laser tuned about 0.1 nm away from the analytical wavelength (the “off-line” measurement). The on-line measurement of the detection was worse than the off-line detection limit when contamination from the analyte was present in the blank.The off-line detection limit was determined as follows. Fluores- cence was measured for various concentrations of the analyte Right electrode Sample port / Fluorescence excitation Left electrode Furnace tube Fig. 1. Modified Perkin-Elmer HGA-500 atomiser. Two 5-mm holes were drilled through the right electrode and two 4-mm holes were drilled through the furnace tube to allow the laser radiation to pass transversely through the furnace assembly. Fluorescence was then observed through the bore of the tube. at the analytical wavelength. The blank was also measured at the analytical wavelength and the signal for the blank was subtracted from the fluorescence signals. The laser was then tuned off-line and 16 measurements of the blank were recorded.The instrumental detection limit was then three standard deviations of the off-line blank. It is possible that at high laser powers and under saturation conditions, fluores- cence from the analyte could be excited with the laser tuned 0.1 nrn off-line, due to saturation broadening.10 However, for the laser powers used here, fluorescence from the analyte was insignificant with the laser tuned off-line. Stray light and furnace continuum background were still within the spectral band pass of the monochromator (ca. 4 nm) and constituted the instrumental blank. This procedure for determining the instrumental limit of detection was the same as that used by Bolshov et al. 1 For those elements that had different on-line detection limits from off-line (due to contamination), both detection limits are reported in Table 2 with the “on-line” detection limits appearing in parentheses.Attenuation of the LEAFS signals was necessary to ensure a linear response of the PMT for the construction of calibration graphs. The LEAFS signals were attenuated by reducing the PMT voltage and by inserting calibrated neutral density filters (Ealing, South Natick, MA, USA) between the atomiser and the PMT. Temperature Measurements Temperature measurements of the HGA-500 furnace were made with an optical pyrometer (Ircon, Series 1100, Skokie, IL, USA). The pyrometer was focused on the outside wall of the furnace, near the centre of the tube, and an emittance of 0.7 was assumed. The output of the pyrometer was amplified and recorded with a digital waveform recorder (Epic Instru- ments, Foster City, CAY USA).Results and Discussion Furnace Heating Rates Fig. 2 shows the temperature versus time profiles for furnace tubes with and without laser ports. The temperature of the furnace power supply was set at 2200 “C for 6 s and the maximum heating rate was used after adjusting the optical sensor. (The HGA-500 power supply is equipped with an optical sensor that provides temperature feed-back to allow maximum power heating to any temperature below 2700 “C.) I 1 L 1 I I 0 1 2 3 4 5 6 7 Time/s Fig. 2. Comparison of heating profiles of a standard HGA-500 tube (A) and a tube with laser orts for ETA-LEAFS (B) (see also Fig. 1). The tube with laser ports feats and cools faster than the standard tube when the same power supply settings are used (i.e.2200 “C for 6 s)JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 431 1.0 I I I I 1 I I 1300 1400 1500 1600 1700 1800 Tern peratu re/"C Fig. 3. LEAFS of lead as a function of atomisation temperature. Non-resonance LEAFS signals for 0.2-ng samples of lead (283 1im/403 nm) were measured at 100 "C increments in the atomisation temperature 108 107 106 a) 105 0 E 2 104 rc a) .- c z 103 a 1 02 10' TI I I I 1 2 2 x 102 2 x 104 2 x 10' Amount of analyte/pg Fig. 4. Linear dynamic ranges for cobalt and thallium. LDR for thallium was obtained for transverse and axial laser excitation and that for cobalt was obtained for transverse excitation only The heating rates for the two tubes were 1400 "C s-1 for the unmodified tube and 1700 "C s-1 for the tube with laser ports.The heating rates were calculated from the time required to reach 2200 "C from room temperature. Because of its lower mass, the tube with laser ports (represented by the solid line, B, in Fig. 2) heated more rapidly than the unmodified tube. During the first 2 s of the atomisation cycle, the tube with laser ports heated to a temperature slightly higher (by about 50 "C) than the programmed temperature. It then cooled to a temperature slightly lower than the programmed temperature before reaching a relatively constant temperature. The tube with laser ports also cooled more rapidly than the unmodifed tube after the atomisation cycle was terminated (after 6 s in Fig. 2). Adjustments of the programmed atomisation tem- perature and the optical sensor temperature could not completely compensate for the anomolous heating and cooling of the tube with laser ports.The different heating characteristics of this tube probably did not affect the LEAFS signal size and detection limit because the over-all temperat- ure difference between the unmodified tube and the tube with laser ports was small (k 100 "C for the same power setting) and small variations in the atomisation temperature did not affect the magnitude of the LbAFS signals. For example, fluores- cence signals for 0.2 ng of lead in the tube with laser ports showed only a slight temperature dependence between 1300 and 1800 "C (Fig. 3). The optimum atomisation temperature for lead was 1600 "C (Table 2). Above 1600 "C furnace emission increased the blank signal and thus degraded the detection limit (see later discussion on furnace emission).Furnace Configurations A standard HGA-500 furnace is 28 mm long with an inner diameter of 6 mm. There are two possible axes for laser excitation. The laser can be directed down the bore of the tube and fluorescence can be collected through a hole drilled in the side of the tube (axial excitation). Alternatively, laser radiation can pass transversely through the tube and fluores- cence can be collected through the bore of the tube (transverse excitation, Fig. 1). Dittrich and Stark8 investigated LEAFS, using both axial and transverse laser excitation, in an atomiser with the same dimensions as the one we used. They chose to machine two slits, which were 4 mm in length and 1.5 mm wide, into the side of their tubes.We drilled two 4-mm circular holes in the HGA-500 tube furnace. Dittrich and Stark8 reported a pre-filter effect with axial excitation. However, no data was shown to confirm this phenomenon. We detected LEAFS for thallium (377 nm/535 nm) and cobalt (304 nm/341 nm) using axial and transverse excitation. (Here, non-reson- ance transitions are expressed as: laser excitation wavelength/ fluorescence observation wavelength.) We could find no difference in detection limit between the two modes of excitation for each element. Our detection limit for thallium in both configurations was 0.1 pg. The limits of detection for cobalt were 0.7 and 0.5 pg for transverse and axial excitation, respectively. The linear dynamic range (LDR) for thallium extended six orders of magnitude above the detection limit for both modes of excitation (Fig.4). Pre-filter effects were not present for LEAFS of thallium at concentrations below 2 X l o 5 pg regardless of the orientation of the excitation beam as demonstrated by a relative slope of one for the thallium calibration graph (Fig. 4). The pre-filter effect was insignifi- cant because the absorption transition for thallium (377 nm) was probably saturated throughout the illumination volume. The LDR for cobalt was obtained only for transverse excitation and extended five orders of magnitude above the detection limit (Fig. 4). For cobalt and thallium, we attribute bending of the LEAFS calibration graphs (Fig. 4), at high concentration (above 100 pg ml-I), to self-absorption effects that were explored in detail and will be discussed in a future publication.For the remainder of our investigations we used transverse laser excitation as this was more convenient for the application of Zeeman-effect background correction.9 Measurement of the Blank The limiting noise on the blank for non-resonance ETA- LEAFS, in the absence of contamination from the analyte, was usually caused by blackbody emission from the graphite furnace tube. For non-resonance ETA-LEAFS of silver, stray laser light was the limiting noise on the blank because the excitation and observation wavelengths were only 10 nm apart (328 nm/338 nm). Stray light was also the limiting noise in blank measurements for resonance ETA-LEAFS. We use the term scattered light to refer to light which is scattered by vaporised and atomised particles inside the heated furnace during the atomisation process.Any light that reaches the PMT which is not caused by the atomisation process is defined, here, as stray light (i.e,, the light which is reflected by the walls of the furnace, the optics, etc.) and is considered to be a different phenomenon from scattered light. For aqueous standards, scattered light was not significant.432 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 Furnace Design Considerations To obtain maximum LEAFS sensitivity it is necessary to illuminate and detect the largest possible number of atoms while maintaining saturation of the atomic transition.llJ2 For example, for an atomic transition which was saturated, such as thallium (377 nm/535 nm), the detection limit was improved by a factor of seven by increasing the diameter of the laser excitation port (and the laser beam diameter) from 3 to 4 mm while saturation was maintained.The L'vov platform has been successfully employed in ETA-AAS to delay the vaporisation of the analyte until the gas temperature of the furnace tube is approximately isother- ma1.6 This has been shown to reduce some types of vapour phase interferences. The L'vov platform should produce the same effects in ETA-LEAFS. However, unlike ETA-AAS, ETA-LEAFS in the HGA-500 furnace relies on illumination of only a portion of the tube furnace when using transverse laser excitation. We investigated the ability of the L'vov platform to confine the atoms to the laser illumination volume by testing two different sized platforms.One was a standard Perkin-Elmer solid pyrolytic graphite platform which was 15 mm long and 4 mm wide, the other was a laboratory constructed platform which was made from pyrolytically coated graphite and was 4 mm x 4 mm (i. e., the same length as the laser excitation port in the HGA-500 furnace tube). Both platforms produced detection limits of 0.1 pg for thallium (377 nd535 nm) which was the same as the detection limit obtained for wall atomisation. Standard Perkin-Elmer plat- forms were therefore used because the size of the platform did not affect the detection limit. Detection limits for wall and platform atomisation were similar for all elements investi- gated (see Table 2).Platform atomisation required a slightly higher atomisation temperature (about 200 "C higher) than wall atomisation to obtain the optimum detection limits. Furnace Emission The heated graphite furnace acted as a strong blackbody emitter at temperatures above 1600 "C. The emission from the furnace occurred at the same time that fluorescence was being excited. Furnace emission was more serious for the less volatile elements (Co, Cu and Mn) than for the volatile elements (Ag, In, Pb and T1) because a higher furnace temperature was required for optimum atomisation. It was also more serious for those transitions that were detected in the visible where furnace emission was more intense. For example, blackbody emission from a graphite furnace tube, heated to 2500 "C and focused into an f/3.5 monochromator through a 0.5-mm slit, was ca. 300 times more intense at 510 nm than at 280 nm.For non-resonance ETA-LEAFS the furnace emission was the dominant signal when measuring the blank and it provided the limiting noise at the detection limit. For the HGA-500 tube atomiser, furnace emission was minimised by focusing the ring of furnace emission around the slit of the monochromator. To reduce the amount of furnace emission that reached the photomultiplier tube, the surfaces of the window mountings of the HGA-500 were blackened and a baffle with a 4-mm diameter hole was placed in the window mounting closest to the monochromator. The signal to noise ratio was maximised by optimising the slit width: A slit width of 0.5 mm was optimum for most non-resonance LEAFS.Despite baffling and imaging, blackbody emission from the walls of the heated tube still reached the detected system partly because of the length of the HGA-500 tube furnace. Fluorescence was only collected from the central section of the tube, which was illuminated by laser radiation, while black- body emission from the whole length of the tube could reach the detector. A short tube furnace is therefore more suitable for LEAFS when furnace emission is the limiting noise. For example, detection limits for copper (325 nm/510 nm) in the 28 mm long HGA-500 furnace were ten times worse than those obtained using an 8-mm tube furnace in our laboratory constructed furnace assembly13 because less emission from the 8-mm tube reached the detector. The design considerations and analytical performance of the short tube furnace are discussed elsewhere.13 Resonance Detection Resonance LEAFS signals were measured for two elements, copper (325 nm) and manganese (278.5 nm), because the non-resonance transitions for both elements required detec- tion at wavelengths in the visible region where furnace emission was intense [i.e., Cu (325 nm/510 nm) and Mn (278.5 nm/403 nm)]. Non-resonance LEAFS for copper (325 nm/510 nm) was used by Bolshov et aZ.1 with a cup atomiser. By viewing fluorescence above the atomiser they may have been able to reduce the amount of furnace emission reaching their detec- tion system. With our HGA-500 tube atomiser, the non- resonance LEAFS detection limit for copper was a factor of 30 worse than Bolshov et al.(see Table 2). Our detection limit for resonance LEAFS of copper was 15 times better than our non-resonance LEAFS detection limit and was within a factor of two of Bolshov et aZ. 1 The non-resonance LEAFS transition that we observed for manganese (278.5 nm/403 nm) involves a non-radiatively coupled transition that to our knowledge has not been reported in the analytical LEAFS literature. The resonance LEAFS transition for manganese (279.5 nm) was determined to be 70 times more intense than the non-resonance LEAFS transition. However, the noise on the blank, for resonance LEAFS, was more severe than for non-resonance LEAFS. The net result was that the detection limit for resonance LEAFS was a factor of seven better than the non-resonance detection limit.The blank for resonance LEAFS was caused primarily by stray light while the non-resonance blank was primarily furnace emission. The reduction of stray light in resonance LEAFS required critical alignment of the laser beam to prevent laser radiation, reflected by the walls of the furnace, from entering the detection system. This alignment 5 s H I 9 J Id I Time - Fig. 5. Chart recorder traces for ETA-LEAFS of thallium (signal versus time) compared with the measurement of the blank in the HGASOO furnace. The mass of thallium (2 pg) used to obtain these peaks is a factor of 10 below the detection limit for AAS in the same furnaceJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 433 Table 2. Comparison of detection limits for LEAFS HGA-SOOT Literature LEAFS HGA-500 wall ETA- AASSS Element Ag .. . . co . . . . cu . . . . cu . . . . In . . . . Mn . . . . Mn . . . . Pb , . . . T1 . . . . Wavelength (Ex/Fl*)/ nm 3281338 304/341 3251510 325 4101451 279. 5/403 279.5 2831405 3771535 atomisation Wall/pg temperature/ "C DLS LDRQ 1600 0.01 5 2100 0.5 5 1900 6 4 2200 0.4 4 1600 0.08 5.5 2100 0.6 4 2100 0.09 4 1600 0.007 6 1600 0.1 6 (0.02) (0.7) (0.3) (0.1) Platfordpg Furnace/pg DLS LDRP 0.008 4.5 (0.04) 0.09 5.5 0.6 4 0.08 4 0.01 6 0.1 - (0.1) * Ex/Fl = excitation wavelength/fluorescence wavelength (if different from excitation). t This work, with contamination limited detection limits in parentheses. $ DL = detection limit. § LDR = linear dynamic range. 7 Reference 1. 11 Reference 2.** Reference 11. tt Reference 14. $$ Reference 15, DLs converted to three standard deviation criterion. DL LDRI 0.17 5.5 0.067 6 0.2 fi 7 0.111 - 0.27 - 0.002~ 5 - - - - 0.00211 6.5 DL - - - 1** 0.2** 0.4** 0.02t-t 0.8tt - Flame, p.p.b. HGA-500 ' DLI LDR* * Pg - 0.8 3 5 2 - - 6 10 5 2 - 5 8 5 20 was made easier by increasing the diameter of the laser port in the HGA-500 furnace from 4 to 4.8 mm. A 2-mm iris was also positioned 50 mm from the furnace to direct the laser beam through the laser ports while blocking any aberrations of the beam from reaching the furnace and producing stray light. Detection Limits for ETA-LEAFS The detection limits for ETA-LEAFS (Table 2) are up to three orders of magnitude better than ETA-AASlS for the same elements. Fig. 5 shows some typical chart-recorder tracings for ETA-LEAFS signals in the HGA-500 furnace as a function of time.The four signals on the right of the figure were produced by a mass of thallium (2 pg) that was an order of magnitude above the ETA-LEAFS detection limit and a factor of ten lower than the detection limit reported for ETA-AAS in the HGA-500.15 Detection limits and linear dynamic ranges for LEAFS in the HGA-500 furnace are compared with literature values for LEAFS in cup atomisers and flame atomisers in Table 2. For all elements investigated, the limits of detection for LEAFS in the HGA-500 are better than those reported for flame LEAFS.11J4 The HGA-500 LEAFS data are comparable to LEAFS in a cup atomiser.1.2 The only exception is thallium which is 50 times worse than the value reported by Tilch et aZ.2 Presumably Tilch et al.used a more sensitive non-resonance transition such as that reported by Hohimer and Hargis16 in which LEAFS is excited at 277 nm and detected at 352 and 353 nm. We chose to excite fluorescence at 377 nm for conve- nience because it did not require frequency doubling of our dye laser output. For ETA-LEAFS of lead (283 nd405 nm), the relative standard deviation (RSD), for 20 measurements of 2.0 ng of lead was 5.8% using the AS-40 autosampler and 7.3% with manual pipetting. For 0.2 ng of lead the RSD increased slightly to 10% for 14 measurements using the AS-40 autosampler. Conclusions We have demonstrated femtogram detection limits for ETA- LEAFS of seven elements using a commercial tube atomiser with both resonance and non-resonance detection of fluores- cence.We have measured fluorescence signals that are linear with concentration (relative slope of one) for 4-6 orders of magnitude above the limit of detection. We have been able to use platform atomisation for ETA-LEAFS with no loss in sensitivity or linearity. Based on these results, we feel the critical parameters for maximising sensitivity of ETA-LEAFS include: a rapidly heated atomiser (>lo00 "C s-I), a large illumination volume and careful baffling of blackbody furnace emission. Other work in this laboratory has shown that a short tube furnace (8 mm as opposed to 28 mm) can be used to reduce furnace emission further without losing LEAFS sensitivity.9 Work is now under way to modify the HGA-500 furnace assembly to accommodate shorter tube furnaces.Further improvements in detection limits for ETA-LEAFS are likely to be realised by the use of non-dispersive detection of fluorescence and optimisation of furnace heating rates above 2000 "C s-1. We have not completely characterised the limiting noises at the detection limit for ETA-LEAFS, but an increased repetition rate of the laser will improve the sampling statistics of the transient ETA signals. This may also improve the precision of the measurement and the detection limits. For those elements that could not be saturated with our existing laser system (Mn, Ag and Co), improved detection limits should result from a more powerful laser system. Work is now under way to modify a commercial furnace assembly and magnet system (Zeeman Accessory, Perkin-Elmer) to utilise all the advantages of modern furnace technology in LEAFS including fast heating rates, platform atomisation, matrix modification and reliable background correction.Our objec- tive is to improve instrumentation for ETA-LEAFS to make it more suitable for the analysis of complicated matrices. This work was supported by the National Institutes of Health under grant GM 32002. R.G.M. was supported by a Research Career Development Award from the National Institute of Environmental Health Sciences under grant number ES 00130. Some of the equipment used in this research was purchased under grants from the Research Corporation, the University of Connecticut Research Foundation and the Petroleum Research Fund administered by the American Chemical Society.This work was presented in part at the XI11434 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 Annual FACSS Conference, St. Louis, MO, USA, September 28-October 3,1986 as paper 477, at the 3rd Biennial National Atomic Spectroscopy Symposium in Bristol, UK, July 23-25, 1986, and at the 25th Eastern Analytical Symposium, New York, NY, USA, October 20,1986, as paper 19. R.G.M. also presented parts of this material for the Society of Applied Spectroscopy Tour Speaker Program during April of 1986. 1. 2. 3. 4. 5. 6. References Bolshov, M. A,, Zybin, A. V., and Smirenkina, I. I., Spectrochim. Acta, Part B, 1981, 36, 1143. Tilch, J., Patzold, H. J., Falk, H., and Schmidt, K. P., paper presented at “Analytiktreffen 82,” Neubrandenburg, GDR, November 1982. Goforth, D., and Winefordner, J. D., Anal. Chem., 1986,58, 2598. Bolshov, M. A., Zybin, A. V., Koloshnikov, V. G., Mayorov, I. A., and Smirenkina, I. I., Spectrochim. Acta, Part B, 1986, 41, 487. Reeves, R. D., Patel, B. M., Mollnar, C. J., and Winefordner, J. D., Anal. Chem., 1973,45, 246. Aggett, J., and West, T. S., Anal. Chim. Acta, 1971, 55, 349. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. Slavin, W., and Manning, D. C., Spectrochim. Acta, Part B, 1980,35, 701. Dittrich, K., and Stark, H.-J., J. Anal. At. Spectrom., 1986, 1, 237. Dougherty, J. P., Preli, F. R., Jr., McCaffrey, J. T., Seltzer, M. S., and Michel, R. G., Anal. Chem., 1987, 59, 1112. Hosch, J. W., and Piepmeier, E. H., Appl. Spectrosc., 1978, 32, 444. Weeks, S. J., Haraguchi, H., and Winefordner, J. D., Anal. Chem., 1978,50, 360. Omenetto, N., and Human, H. G. C., Spectrochim. Acta, Part B, 1984, 39, 1333. Preli, F. R., Jr., Dougherty, J. P., and Michel, R. G., Anal. Chem., in the press. Human, H. G. C., Omenetto, N., Cavalli, P., and Rossi, G., Spectrochim. Acta, Part B, 1984,39, 1345. Slavin, W., “Graphite Furnace AAS. A Source Book,” Perkin- Elmer, Norwalk, CT, 1984. Hohimer, J. P., and Hargis, P. J., Jr., Anal. Chim. Acta, 1978, 97, 43. Paper J7f2 Received January 13th, 1987 Accepted March 2nd, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200429

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 435 Determination of Selenium by Graphite Furnace Atomic Absorption Spectrometry Part 2.* Role of Nickel for Analyte Stability Jiii D&dina,t Wolfgang Frech,S Anders Cedergren, lngela Lindberg and Erik Lundberg Department of Analytical Chemistry, University of Umeii, S-901 87 UmeA, Sweden The chemistry involved in the determination of the selenium in the presence of nickel nitrate and nickel chloride was investigated for various furnace conditions and matrix compositions. To obtain a deeper understanding of this complex system, we had to combine information from several techniquessuch as mass spectrometry, radioactive measurements, thermodynamic calculations and AAS measurements using different types of graphite material and furnace designs.It is shown, and discussed, why nickel alone is not sufficient to stabilise selenium if an excess of organic material is present. Furthermore, experimental results obtained by using a two-step constant temperature furnace indicate that the presence of glucose decreases the number of free atoms formed during atomisation. Keywords: Graphite furnace; graphite surface; atomic absorption spectrometry; selenium stabilisation; nickel modifier Nickel is widely used in the determination of selenium by graphite furnace atomic absorption spectrometry (GFAAS)1-4 and it has been shown that this modifier can stabilise selenium up to ca. 1500 K. However, several investigators have reported that the stabilising ability of nickel is, sometimes, partly lost depending on the form of the analyte and the matrix composition.For example, Styriss performed mass spec- trometric measurements on the reaction products from a graphite tube under atmospheric pressure and found that selenium was lost as Se2(g), SeO(g) and SeOz(g) when nickel dichloride was used as a modifier. No losses were observed by this worker when nickel nitrate was used. In agreement with these results, experiments using 75Se as a radiotracers showed that selenium was stable up to 1450 K. However, in the presence of excess of glucose, selenium was unstable and losses commenced during drying. Saeed et al.7 studied the stability of metabolised selenium in blood using a radiotracer and found that selenium was not completely stabilised when using the normally recommended 20 pg of nickel as the nitrate.A five-fold increase in the amount of modifier did, however, prevent these losses. It should be mentioned that a lower sensitivity is to be expected with larger amounts of added nickel.8 In Part 1 of this series,9 the role of the graphite surface on the stability of selenium was discussed. In this paper we will deal with the mechanisms leading to selenium stabilisation during thermal pre-treatment when using nickel salts in various matrices. In order to obtain a deeper understanding of the chemistry involved, we had to combine information from several independent techniques such as mass spectrometry,5 radioactive measurements with 75Se, thermodynamic calcula- tions and AAS measurements using different types of graphite furnaces. Experimental Instrumentation For the atomic absorption measurements, an improved version of the two-step furnace was used.9 The furnace has a * For Part 1 of this series, see reference 9.t Present address: Czechoslovak Academy of Sciences, Institute of Nuclear Biology and Radiochemistry, 142 20 Prague 4, Czechoslovakia. $ To whom correspondence should be addressed. T-shaped configuration incorporating a tube with integrated contacts and a graphite cup below. For the determination of activation energies and appearance temperatures tubes with- out cups were used. A more detailed description of the instrumentation and procedures is given in Part 1 of this series.9 For the radioactive measurements, a Perkin-Elmer HGA-74 furnace was used. The y-radiation was measured with a 45 X 50 mm NaI(T1) well-type scintillation detector.(For details see Part 1 of this series.) For the platform experiments a Varian SpectrAA-30 atomic absorption spec- trometer with a GTA-96 graphite furnace, a DS15 data station and an MX-80 I11 printer were used (see Table 1 for the details of the furnace conditions used). Materials In the Varian SpectrAA-30, pyrolytic graphite coated (PGC) tubes (Varian Part No. 63-100001-00) with solid pyrolytic graphite (Varian Part No. 63-100003-00) or polycrystalline electrographite platforms (EG) were used. The latter were laboratory made from RW07 high-density graphite (Rings- dorff-Werke GmbH, Bonn-Bad Godesberg, FRG) to have the same surface area and mass as the pyrolytic graphite platforms.High-temperature Equilibrium Calculations The calculations were performed as described in Part 1 of this series.9 Results and Discussion Stabilisation of Selenium in Aqueous Solutions Using Nickel Nitrate Based on mass spectrometric observations, Styriss proposed that nickel forms stable NiSe(s) through the reaction between Se02(s) and Ni(s). This suggestion is in line with the calculations given in Fig. 1. It should be recalled that the suggestions for obtaining accurate results arelo that all relevant compounds of the system are considered, that the thermodynamic data are correct and that reactions are fast enough (except for the heterogeneous reactions between carbon and oxygen) to attain a state close to equilibrium. The right-hand diagrams (a’+’) show the distribution of the nickel species as a function of temperature for different partial pressures of oxygen, while corresponding selenium species are436 0.01 5- - 0.010- 5.'r- 0.005 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 - , - . * ., 0 NiO(s) - ( a ' ) 0,015 - 0.010 5. -x 0.005 500 700 900 r . , . v - NiO(s) - - ( b ' ) 0 ' Table 1. Furnace conditions for the Varian SpectrAA-30 Gas Ramp Hold Temperature/ flow-rate/ Stage time/s time/s K ml min-1 Drying . . . . 10 20 483 300 Pre- trea tmen t 20 40 Varied 300 Cool down step 15 5.0 573 300 Atomisation . . 1.2 5 .O 2173 0 Cleaning . . 1.0 4.0 2773 300 Wavelength . . . . 196.0nm Samplevolume . . . . lop1 Spectral band width . . 1.0 pm Lampsource . . . . EDL,7W Modifiervolume . . . . 5pl Table 2.Recovery of 2 ng selenium(1V) in the presence of nickel as nitrate after thermal pre-treatment using Ar or H2 as sheath gases. The pre-treatment temperature was 1173 K. PGC tubes and pyrolytic graphite platforms were used Sheath gas Mass of Ni/pg Recovery, % . . . . . . . . Ar 0.5 64 5 100 H2 0.5 57 5 91 . . . . . . . . I I I I I ] 500 700 900 700 900 Temperature/K Temperatu re/K Table 3. Activation energies and appearance temperatures for 5 ng of selenium stabilised with 5 pg of nickel as nitrate using a PGC tube Fig. 1. Distribution of selenium and nickel species as a function of temperature for three partial ressures of oxy en: (a and a ' ) , pOz 10-6 bar; (b and b ' ) , p 0 2 lO-249ar; and (c and c 3 , equilibrium. For (c and c') even oxygen is assumed to be in equilibrium with carbon.In addition to the input amounts and species given in Part 19 we used 0.017 pmol of nickel as input amount and considered Ni( , Ni(C0)4 g), NiC12( ), Ni(l), Ni(s), NiC12 s , NiO(s), Ni3C($ NiSe03(!), NiSel,05fs), NiSel,14(s), NiSel.,Js] and NiSe2(s). Thl input amount of hydrogen was 0.056 pmol given by the left-hand figures (u-c). In accordance with the arguments given in Part 1 of this series,g the partial pressure of oxygen in the graphite tube is likely to decrease by several orders of magnitude from the bulk gas phase to the region which is in direct contact with a reactive surface. At temperatures below 600 K the reactivity of the graphite surface is normally low and as a consequence of this it is reasonable to assume that the oxygen pressure is sufficiently high to stabilise selenium in the form of NiSe03(s).For higher temperatures and lower partial pressures of oxygen, various condensed non-stoicheiometric nickel selenides will be for- med, according to these calculations. It is well known that solid nickel oxide can be reduced to solid nickel at low temperatures by a mixture of carbon monoxide and hydrogen. The presence of solid nickel seems to be essential for selenium stabilisation. Results reported by Esser and Durnberger" show that for the direct determination of selenium in inorganic solid materials, losses occur when nickel is added in the form of nickel oxide. If the solid sample is thoroughly mixed with an excess of nickel powder, however, the conditions for the determinations are improved.As can be seen in the left-hand diagrams of Fig. 1 (a-c), under reducing conditions significant fractions of the volatile selenium compounds Se2 and H2Se are expected to be formed at elevated temperatures. It should be pointed out that for 900 K no partial pressure of oxygen was found for which a fraction of condensed selenium close to 100% was obtained. Results from radiotracer experiments with 75Se, as well as MS Appearance Pre-treatment Gas EJkJ mol- 1 temperature/K temperature/K Ar . . . . 255k27 1473 k 24 1123 Ar . . . . 277k7 1473 f 5 753 CO . . 158f24 1263 k 60 753 H2* . . 186k7 1263 f 10 753 * 50 ng Se. measurements by Styris,s showed no losses in the temperature interval considered in Fig. 1. This can be explained by slow reactions between hydrogen or oxygen and the solid nickel selenide.Results obtained at a higher temperature, 1173 K, are summarised in Table 2. It is interesting to note that a ten-fold increase in the amount of nickel as nitrate prevents losses in an argon but not in a hydrogen atmosphere. As discussed elsewhere,lO hydrogen is always formed in a graphite tube and this may be the reason why losses due to the formation of H2Se(g) are seen for 0.5 pg of nickel as nitrate in argon. For the larger amount of nickel, less hydrogen will be available for reaction with selenium because of the oxidising character of this modifier. In separate radiotracer experiments with carbon monoxide instead of hydrogen, no losses were observed at 1173 K for 5 pg of nickel as nitrate. From these results we conclude that the postulated condensed nickel selenides are stable even under the probably reducing conditions prevailing in a carbon monoxide atmosphere. In mass spectrometric vacuum experiments,s a nickel selenide sample was found to be stable up to 1150 K.The results shown in Table 3 were obtained as described in Part l , 9 i.e., by means of wall atomisation using spatially isothermal side-heated tubes. The fact that the activation energy values in argon are not changed by the different pre-treatment temperatures given in the table indicates that the selenium atoms are formed from the same precursor. The reason for the lower energy values for carbon monoxide and hydrogen cannot be explained.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 437 The fact that the surface upon which the sample is deposited is important for the stabilising capability of nickel nitrate is shown by the results given in Table 4.Even when the matrix was pre-treated at high temperatures to pyrolyse the glucose completely, a lower recovery of selenium was observed in some experiments. This means that glucose is likely to change the nature of the surface to either prohibit the formation of "free" solid nickel or to alter the form of selenium such that the stabilisation is hindered. The poor reproducibility in between the different measuring series reflects, in our opinion, insufficient control over the condition of the graphite surfaces of the different platforms/tubes used in these experi- ments. As was discussed in Part 1 , 9 carbon monoxide as well as hydrogen will affect the structure of the surface. The same is, of course, also true for gaseous reaction products generated from substances such as glucose.In order to investigate the extent to which selenium is retained at various temperatures in PGC and EG tubes the following experiments were performed with the two-step furnace which permits separate heating of a graphite cup and tube. The cup is positioned beneath the tube and the volatilised sample diffuses from the cup into the tube (see Part 1). During a first firing the cup was heated to 2373 K while the tube was kept at the various temperatures shown in Table 5. The fraction of selenium that was retained in the tube was estimated from a second firing during which both tube and cup were heated to 2373 K.The specific signals observed during the first and second firings are reported in Table 5. It should be mentioned that retention values are uncertain due to the fact that retained selenium is volatilised and atomised in the course of the rising tube temperature. Even at tube temperatures as high as 1723 K, complete retention of selenium was observed for polycrystalline graphite as compared with almost no Table 4. Recovery (YO) of 2 ng of 75SeIV in the presence of 20 pg of nickel as nitrate using a pre-treatment temperature of 993 K. Before sample injection, 200 pg of glucose were pre-treated at various temperatures for 60 s. Measurements were repeated on different occasions (A-C) using new sets of graphite tubes and platforms.PGC tubes and pyrolytic graphite platforms were used Pre-treatment temperature/K Measurement 483 813 993 A . . . . . . 63f3 80+8 83k9 B . . . . . . 72213 98k4 10022 C . . . . . . 86f9 98f2 93+4 retention whatsoever in PGC tubes. This indicates that secondary surface reactions involving nickel and selenium should take place at the cooler parts of spatially non-isother- ma1 end-heated graphite tubes, and should be more pro- nounced in polycrystalline graphite tubes. A comparison with corresponding experiments in the absence of nickel (see Part 1) shows that the retention of selenium species is signficantly smaller. Stabilisation of Selenium in Chloride-containing Organic Matrices Using Nickel Nitrate In order to understand the nature of the volatilsation interference effects obtained in the presence of glucose and sodium chloride, we performed radiotracer recovery measure- ments using a dual-cavity platform.12 As can be seen in Fig.2, the largest losses of selenium were obtained when all components were placed in the same cavity. The lower recovery (as compared with the upper curve) observed when sodium chloride and glucose were placed in separate cavities must be ascribed to effects caused by gaseous decomposition products from the glucose - sodium chloride mixture, such as furan derivatives, hydrocarbons ,*3 hydrogen and carbon monoxide. Selenium often has to be determined in chloride-containing matrices. In order to understand the effect of chloride on selenium stability, a number of equilibrium calculations were performed on the C, H, 0, Se, Ni and Cl system.Fig. 3 shows the results obtained for 0 and 28 nmol of hydrogen, A loot .- ' 5 60 LSe + Ni(NOJ2 I + I \ \ + NaCl + glucose a 4 0 1 d!m w\ 2o t ! Se + Ni(N0312 I NaCl + glucose I I O ' 5b0 1000 1500 2( Tern peratu re/K 30 Fig. 2. Dual-cavity platform recovery study for 2 ng of '%eIV using 20 pg of nickel as nitrate as modifier and a mixture of 200 pg of sodium chloride and 400 pg of glucose as matrix. A, No matrix present; B, matrix and analyte plus modifier placed in separate cavities; C, matrix and analyte plus modifier mixed Table 5. Fraction of selenium volatilised/atomised and retained in the presence of 1 pg of nickel as nitrate at various tube temperatures in PGC and polycrystalline electrographite (EG) tubes.Cup temperature was 2373 K throughout; 2 or 5 ng of selenium(1V) were used Fraction of selenium atomised in the first firing,* YO Fraction of selenium retained , t O h Tube temperature/K EG tube PGC tube EG tube PGC tube Tube unheated$ <2 <2 101 t5 116t 12 1573 15+5 30 f 15 125+5 62+ 10 1723 17 f 5 90 t 5 137f5 3 f 2 <2 1823 3+2 - 1923 103 t 5 1993 100 + 5 98 k 5 <2 <2 - 95 + 5 - - * Related to 2373 K using the correction factor (T/2373)l-YS. t Estimated from the peak area of a subsequent firing at tube and cup temperature of 2373 K following the firing (with injected sample) at the $ A temperature of 773 K is assumed for the diffusion correction. specified tube temperature.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 I I I 1 r 1 500 700 900 500 700 900 Temperatu re/K Temperature/K Fig.3. Distribution of selenium and nickel species in the presence of chlorides as a function of temperature without and with hydrogen: (a and a’), no hydrogen; and ( b and 6’ ,28 nmol hydrogen. Oxygen was chlorine was 0.034 vmol. Other input amounts and species considered as in Fig. 1 assumed to be in equilibrium wit h carbon. The input amount of 0 - *-GPO 0% HZ 0.1% H2 500 600 700 Temperature/K Fig. 4. Calculated fraction of volatile selenium compounds in the presence of nickel (0.017 pmol as dichloride for various hydrogen - argon mixtures as a function o f’ temperature respectively. For simplicity, only results corresponding to equilibrium between carbon and oxygen are represented. In the absence of hydrogen, selenium is thermodynamically not stable (in the presence of nickel chloride), in contrast to the results obtained for selenium in the presence of nickel as the nitrate.It should be observed that in the absence of hydrogen no solid nickel is present in a nickel chloride matrix in contrast to nickel nitrate. In the presence of hydrogen, condensed nickel is, however, present at 500 K in a nickel chloride matrix and the main selenium compound is expected to be NiSel.05(s). According to mass spectrometric measurements under atmospheric pressure, selenium was found to be partially lost between 500 and 1000 K in the form of Se2(g), SeO(g) and Se02(g).5 For these experiments, 50 ng of selenium and 200 ng of nickel as the dichloride were used. On using the same amounts of selenium and nickel dichloride or nitrate, we did not observe any significant losses of selenium below 1000 K in our radiotracer experiments. The reason for the losses observed by Styris in the presence of nickel dichloride could be ascribed to a relatively smaller amount of solid nickel, formed during pre-treatment as compared with the situation when nickel nitrate was used.The importance of 0.8 0.6 0.4 0.2 0 0.6 0, c m 2 0.4 2 a I) 0.2 0 0.4 0.2 C H2 .’ Tube temp. 2123 K r . . I B : : (0.55): ; Cup temp. 2123 K - _ _ _ - - ’ * / P co /---- / C / I I I I I I #------ Ar / C 0 2 4 6 8 10 Ti me/s Fig. 5. Absorbance versus time profiles for selenium in different sheath gases (hydrogen, carbon monoxide and argon) using the two-step atomiser. The broken lines (C) indicate the temperature versus time profiles of the cup.5 pg of nickel as chloride (A) or nitrate (B) were used. Peak areas (A s) indicated in parentheses hydrogen for the stability of selenium is further illustrated in Fig. 4. As can be seen, small amounts of hydrogen are necessary for stabilisation, but large amounts lead to increased volatility of selenium. This effect is much more pronounced at higher temperatures. Gas Phase Interference Effects Fig. 5 summarises results obtained with the two-step furnace using hydrogen, carbon monoxide and argon as sheath gases and nickel nitrate or nickel chloride as modifiers. Considering the results for nickel nitrate, the slightly lower peak-area value obtained in the carbon monoxide atmosphere as compared with argon can be ascribed to the higher rate of atom removal in the former medium. In separate experiments with lead and cadmium the sensitivity quotients between argon and carbon monoxide were in the range 1.0-1.2.The corresponding quotient for argon - hydrogen was 6.0-6.2. For selenium this value was 13 which means that the atomisation efficiency is a factor of two lower in hydrogen. It should be emphasised that no losses of selenium at low temperatures occur. If the tube temperature was increased to 2500 K, the peak-area sensitivityJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 439 0.4 in 4 0.2 : I-+ PG platform . ' EG platform \\, I \ 1 \ I I I 0 4 / 6AO 1000 1400 1800 Tern pe ra t u re/K Fig. 6. Thermal pre-treatment curves for 2 ng of selenium in the presence of 20 pg of nickel as nitrate using a pyrolytic raphite (PG) (broken line) and polycrystalline electrographite (EGf platform.A Varian Techtron GTA-96 graphite furnace under gas stop conditions was used Tube temp. 2123 K 0.6 01 S (D 0.4 e s 2 11 0 --- /-/ I / ' A I r\, / I / B ' - -- D 0 3 6 9 12 Time/s Fig. 7 . Absorbance versus time profiles for 5 ng of selenium(1V) in the presence of 5 pg of nickel as nitrate and various matrices using the two-step furnace: A, 5 pg Ni (peak area 0.74 A s); B, 50 pg NaCI + 100 pg glucose + 5 pg Ni (peak area 0.40 A s); and C, 5 pg Ni + 100 pg glucose (peak area 0.59 A s). The broken line (D) indicates the temperature versus time profile of the cup ratio decreased to six. This is in good agreement with theoretical predictions, namely that the formation of hydrogen selenide becomes insignificant at 2500 K.As can be seen for the experiments using nickel chloride, the peak-area sensitivity is in all gases lower than the corresponding values for nickel nitrate. However, this signal decrease is much more pronounced in the carbon monoxide atmosphere. As gaseous selenium chlorides are rather unstable at these high tempera- tures,lO this effect could be a result of an increase in the partial presure of carbon in the carbon monoxide environment. Equilibrium calculations show that an increase in the partial pressure of carbon from the equilibrium pressure of 2 x 10-10 bar at 2100 K to 5 x 10-9 bar results in an enhancement in the formation of gaseous selenium carbide of from 2 to 50%.It is realistic to assume that the partial pressure of carbon is lower when nickel nitrate is present as compared with nickel chloride. As can be seen in Fig. 6 the peak-area sensitivity is significantly lower when using a polycrystalline graphite platform compared with a pyrolytic graphite platform. It should be mentioned that when graphite powder was added to a pyrolytic graphite platform the peak-area absorbance signal decreased by 20%. The decrease in sensitivity observed can be ascribed to an increase in the formation of hydrogen selenide as more hydrogen is expected to be formed in the presence of polycrystalline graphite.14 Fig. 7 shows that glucose reduces the atomisation efficiency for selenium. The latter effect could be a result of the formation of gaseous selenium carbide or hydride in light of the arguments given above.In separate experiments we found that the effect of glucose alone or mixed with sodium chloride on the peak-area sensitivity was insignifi- cant if the atomisation temperature of the two-step furnace was increased to 2473 K. Conclusions The chemistry involved in the determination of selenium in the presence of both nickel and matrices containing organic compounds is difficult to control as the reactions leading to stabilisation of selenium are very complex. In addition, the presence of organic substances can lower the atomisation efficiency owing to the formation of selenium molecules during the atomisation step. The trapping of volatile selenium on heated graphite15 should be more efficient in graphite tubes which have been treated with nickel.We would like to thank B. Hiitsch, Ringsdorff-Werke GmbH, for supplying graphite parts and Dr. J. Moore of Varian Techtron Pty. Ltd., for providing the SpectrAA-30 and GTA-96 on loan. This work was supported by grants from the Swedish Natural Science Research Council. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Ediger, R. D., At. Absorpt. Newsl., 1975, 14, 127. Norheim, G . , Saeed, K., and Thomassen, Y., At. Spectrosc., 1983,4, 99. Carnrick, G. R., Manning, D. C., and Slavin, W., Analyst, 1983, 108, 1297. Saeed, K . , and Thomassen, Y., Anal. Chim. Acta, 1982,143, 223. Styris, D. L., Fresenius 2. Anal. Chem., 1986,323, 710. Cedergren, A., Lindberg, I., Lundberg, E., Baxter, D. C., and Frech, W., Anal. Chim. Acta, 1986, 180,373. Saeed, K., Thomassen, Y. , and Langmyhr, F. J., Anal. Chim. Acta, 1979, 110,285. Welz, B . , Schlemmer, G., and Vollkopf, V., Spectrochim. Acta, Part B, 1984, 39,501. Dedina, J., Frech, W., Lindberg, I., Lundberg, E., and Cedergren, A., J. Anal. At. Spectrom., 1987, 2 , 287. Frech, W., Lundberg, E., and Cedergren, A., Prog. Anal. At. Spectrosc., 1985, 8, 193. Esser, P., and Durnberger, R., Lecture presented at 2nd International Colloquim on Solid Sampling with Atomic Spectroscopic Methods, Wetzlar, FRG, 1986. Welz, B., Akman, S., and Schlernmer, G., Analyst, 1985,110, 459. Prey, V., and Gruber, H., Stiirke, 1977,4, 135. Frech, W., and Cedergren, A., Anal. Chim. Acta, 1976,82,93. Sturgeon, R. E., Willie, N. S. , and Berman, S. S., Fresenius 2. Anal. Chem., 1986, 323, 788. Note-Reference 9 is to Part 1 of this series. Paper J7l15 Received February 2nd, 1987 Accepted April 13th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200435

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 441 Lead Atomisation from Soil by Slurry Introduction Electrothermal Atomisation Atomic Absorption Spectrometry Part I. Effects of Matrix Components on the Absorbance versus Time Profile Michael W. Hinds and Kenneth W. Jackson* Department of Chemistry, University of Saskatchewan, Saskatoon, Saskatchewan S7N OWO, Canada A typical soil matrix was modelled, in order to examine the effects of matrix components on the absorbance signal for Pb using L'vov platform atomisation. Components studied were clay (montmorillonite), organic carbon (as humic acid), phosphate, Mg2+ and CI-. Lead is adsorbed on the clay particles, and this delayed the signal to a higher temperature compared with an aqueous Pb solution. Phosphate and Mg*+ also delayed atomisation, but not a t the low concentrations normally found in soil.There was no CI- interference at normal concentrations in soil. Organic carbon caused Pb to atomise at a lower temperature. This could lead to a systematic error, compared with aqueous standards, if atomisation occurs prior to conditions being nearly isothermal in the graphite furnace. The use of a matrix modifier is recommended to overcome this interference. Keywords : Electrothermal atomisation atomic absorption spectrometry; slurry; lead atomisation; matrix effects; soil analysis It was shown previously that slurry introduction electrother- mal atomisation atomic absorption spectrometry (slurry - ETA-AAS) is rapid and convenient for determining Pbl and Cd1J in soil.The technique is much simpler than conventional acid digestion flame atomic absorption spectrometry, and it has several advantages compared with solid sampling proce- dures which involve directly weighing solid aliquots into the electrothermal atomiser .3 These include easier sample hand- ling (because slurry aliquots are micropipetted into the electrothermal atomiser) and less risk of errors due to sample inhomogeneity. 1 Although slurry - ETA-AAS using L'vov platform atomisation and aqueous calibration standards gave accurate recoveries of Pb and Cd,lJ subsequent work in our laboratory has shown differences in Pb absorbance peak shapes and appearance times between many soils. In some instances, the signals are delayed with respect to aqueous Pb standards. This may not be a problem if standards and samples are atomised under near-isothermal conditions and absor- bance peak areas are measured.4 However, soils high in organic matter often give Pb signals that are early shifted with respect to aqueous Pb. This can result in loss of Pb if the sample is being vaporised before the electrothermal atomiser has attained near-isothermal conditions; i.e., during the period of rapid temperature gradient as the graphite tube is being heated. Losses can also occur during the ETA charring stage. In order to avoid the risk of these losses happening, the soil matrix requires some treatment. This could be whole or partial digestion prior to slurry preparation, or matrix modification during the ETA-AAS determination, Matrix modification is preferred, but it requires at least a partial understanding of the factors affecting Pb atomisation from soil particles.Slurry - ETA-AAS has been applied to a variety of sample types, including Ti02,5 soils,lJ coal,6 sewage sludge7 and foods.8.9 However, most of the work has been empirical in nature. Little has been done to attempt to understand the fundamental atomisation processes that are taking place, which may be quite different from the atomisation of samples introduced as solutions. Recent work from this laboratorylOJ1 has shown that metals either occluded within or adsorbed on * Present address: Wadsworth Center for Laboratories and Research, New York State Department of Health, Empire State Plaza, Albany, NY 12201, USA. to alumina particles have atomisation characteristics that are different from the metal in aqueous solution.In order to extend these fundamental studies to soils, a perceived difficulty was the physicochemical complexity of the soil matrix, which would make it difficult to identify the effect of any individual matrix component on the atomisation charac- teristics. Soil is a dynamic mixture of degradation products of rocks and minerals, also containing 0-30% organic matter from biological decomposition. 12 The inorganic material is dominated by oxides of Si, A1 and Fe, with lesser amounts of carbonates (of Ca and Mg), sulphates, halides, phosphates and nitrates.13 Trace metals can be distributed between most of the soil components depending upon the components, pH and moisture c0ntent.14~15 Lead, when present as a pollutant, is adsorbed on to the clay particles of soil and/or bound by organic matter coating the particles.13 In order to study the effects of selected soil matrix components on Pb atomisation, the matrix was simplified by preparing an artificial soil.A clay mineral (montmorillonite) formed the basic matrix and, from prior knowledge of ETA-AAS interferences, other matrix components which might affect the atomisation characteristics of Pb were added. These were organic carbon (as humic acid), Mg2+, phosphate and C1-. Other potential interferents are unlikely to be present in high enough concentrations to have any effect. A similar approach was used by Guy et aZ.16 for studying the extraction of trace metals from lake sediments. They made an artificial sediment based on a clay mineral (bentonite), MnO and humic acid.The purpose of the present study was to examine the individual and combined effects of the selected matrix components on the Pb absorbance peak shapes and appearance times. This would indicate whether matrix modification was needed to delay the onset of Pb atomisation during the quantitative analysis of some soils. Experimental Apparatus A Perkin-Elmer Model 2280 atomic absorption spectrometer with deuterium-arc background correction, and a Model HGA-500 electrothermal atomiser were used. The graphite tube wall temperature was monitored by focusing an optical pyrometer (Series 1100, Ircon, Niles, IL, USA) on the outside wall of the tube. As described previously17 the spectrometer and pyrometer were interfaced to a 12-bit A/D board (Model442 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 DT2801, Data Translation, Marlborough, MA, USA) mounted in an IBM-PC microcomputer. This enabled rapid (1/60 s) simultaneous monitoring of absorbance and temperature and high-speed background correction.Perkin-Elmer pyrolytic graphite coated tubes and pyrolytic graphite (L'vov) platforms were used. All absorbance measurements were made at 283.3 nm, and background- corrected absorbance peak areas (A s) were measured. The electrothermal atomiser operating parameters are given in Table 1. Samples and Reagents Soil samples were typical A horizon Saskatchewan soils currently used in agricultural production. The montmorillo- nite (Ward's Natural Science, Mississauga, Ontario, Canada) from Upton, Wyoming was described in the American Petroleum Institute Research Report No. 49.18 All soils and montmorillonite were dried at 105 "C for 24 h and ground to <20 pm particle diameter using a vibratory mill (Retsch Model MM2, Brinkmann Instruments, Rexdale, Ontario, Canada).Previous work2 showed that atomisation efficiency is comparable to solutions provided slurry particles are <20 pm in diameter. Analytical-reagent grade chemicals were used throughout. No measureable amounts of Pb were present in the distilled water. Standard stock Pb solution (1000 mg 1-1) was prepared by dissolving Pb(N03)2 (Analchemia Chemicals, Mississauga, Ontario, Canada) in 1% HN03. Calibration standards were made by dilution of the stock solution.Stock solutions (10%) of Mg(N03)2.6H20 (BDH Chemicals, Toronto, Ontario, Canada) and (NH4)2HP04 (BDH Chemicals) were prepared. A solution (0.05 mg 1-1) of lead(I1) cyclohexane butyrate (Alpha Products, Danver, MA, USA) in xylene was also prepared. The sodium salt of humic acid (Aldrich Chemicals, Milwaukee, WI, USA) was acidified by mixing with 1% HN03, allowing to stand for 0.5 h and discarding the supernatant. A stock humic acid slurry was made by stirring overnight 0.12 g of the acidified material in 80 ml of water. Procedure Slurries were prepared by adding exactly 20 ml of distilled water to 60 mg of the finely ground soil or 30 mg of montmorillonite in a beaker and stirring magnetically for a minimum of 5 min. Appropriate amounts of additives, such as humic acid, Mg(N03)2 or (NH4)2HP04 were introduced at this time.Slurry aliquots (20 pl for the soil or 10 p1 for the montmorillonite) were then injected with an Eppendorf micropipette into the furnace after an additional 5 min of stirring. All montmorillonite slurry aliquots contained 0.63 ng Pb and 15 pg of solid matrix. Soil slurry aliquots contained 0.62-0.70 ng Pb and 60 pg of solid matrix. The operating parameters in Table 1 were used for all samples, but for aqueous Pb solutions without a matrix modifier, care was needed not to exceed the charring temperature of 900 "C. Otherwise, volatilisation losses of Pb could occur. ~~ ~ ~ Table 1. Instrumental operating parameters ETA stage Parameter Dry Char Atomise Clean Holdtimek . . . . 60 30 5 5 TemperaturePC .. . . 130 900 1900 2700 Ramptimeh . . . . 10 20 0 1 Wavelength . . . . 283.3nm Spectral band width . . 0.7 nrn Lampcurrent . . . . 8mA Purgegas . . . . . . N,(flowstopped during the atomise stage) Results and Discussion Atomisation of Lead from the Clay Matrix In order to study this atomisation behaviour, the aluminosili- cate clay montmorillonite was used. This material has a defined crystal structure and the chemical compositions of its forms are known.18 It was analysed for Pb by ETA-AAS following complete digestion with a mixture of HN03 - HC104 - HF.2 The Pb concentration in the montmorillonite was 42 L- 2 yg 8-1. It was necessary to confirm whether this Pb was associated with montmorillonite in a similar way to the Pb in soil; i. e., if it was adsorbed on the particles.In a soil slurry, the adsorptive forces are strong enough to prevent Pb from leaching into the aqueous phase. However, Agemain and Chau19 found that mild acid extraction would remove adsorbed metals without seriously degrading the silicate matrix; i.e., it would not extract occluded metals to a large extent. Therefore, two leaching tests were carried out. First, an aqueous montmorillonite slurry was centrifuged and analysis of the supernatant by ETA-AAS confirmed that pure water had not leached out any measurable amount of Pb. Second, a method to remove surface adsorbed Pb, described by Guy et al. 16 was used. Dilute HN03 (pH = 1) was added to montmorillonite, the mixture was stirred, centrifuged and the clear colourless supernatant was found to have a high Pb concentration.The remaining solid was slurried with distilled water and slurry - ETA-AAS confirmed that there was no longer any measurable amount of Pb associated with the particles. Hence, the mild acid extractant had removed all of the Pb. Lead was re-introduced to the remaining solid by adding 20 ml of a 0.05 mg 1-1 solution (without pH adjustment), stirring for 5 min, and analysing again by slurry - ETA-AAS. The appearance time (tapp) and shape of the resulting Pb absorbance profile closely matched the signal profile prior to extraction, indicating Pb was re-adsorbed. Also, the Pb absorbance preceded any non-specific absorbance. Work with Pb occluded within alumina11 showed that tapp was much later; i.e., the Pb signal coincided with the non-specific absorbance associated with matrix break-up.Hence, the above tests confirmed that all of the Pb associated with montmorillonite is adsorbed. It was shown by Karwowska and Jackson11 that the release of Pb from alumina matrices during slurry - ETA-AAS was delayed compared with aqueous Pb solutions. Kinetic studies indicated that the delay was caused by adsorption of Pb on the alumina particles. The L'vov platform absorbance peak for the aqueous montmorillonite slurry is shown in Fig. 1. The peak is delayed compared with the aqueous Pb standard, so adsorption on the clay matrix has a similar effect to adsorption 1.0 I 1 1 I \ \ 1 0 1 2 3 4 5 Tim e/s Fig. 1. Effect of montmorillonite on the Pb absorbance signal: A, an aqueous Pb solution (0.75 n$ Pb); B, a montmorillonite slurry (0.63 ng Pb and 15 pg montmorillonite); and C, the tube-wall temperatureJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 443 - - on alumina in retarding analyte vaporisation. It is not clear from these results whether the adsorption of Pb on montmoril- lonite is chemical or physical. K 2 1600 3 0) n 14OOk 6 Effect of Organic Carbon In conventional solution analysis by ETA-AAS, organic material can cause an early shift of the Pb absorbance peaks,20721 and it has been suggested20 that losses may occur from the formation of volatile Pb compounds during the charring step. Both of these effects are seen in Fig. 2, which shows the absorbance signal for 1 ng Pb introduced as an organolead compound. Both tapp and tpeak (the peak maxi- mum time) are shorter, and the peak area is smaller, compared with an aqueous standard containing the same amount of Pb.An early shift is also seen with many soil slurries. Fig. 3(a) shows slurry - ETA-AAS signals for two soils, one with 4.7% organic carbon (soil 8) and the other with 0.3% organic carbon (soil 7102). (Organic carbon was determined by the procedure described in Page et aZ.22) Compared with the aqueous Pb standard, soil 7102 has a delayed absorbance, but for soil 8 tapp is similar to the aqueous standard. This is consistent with the organic carbon offsetting the delay in absorbance caused by other soil matrix effects. In order to confirm this, the organic carbon was removed by ashing both soils at 600 "C for 24 h in a muffle furnace.They were then re-analysed by slurry - ETA-AAS, and both absorbance signals were delayed compared with the aqueous standard [Fig. 3(b)]. Integrated absorbance measurements indicated no losses of Pb during this ashing treatment. Humic acid was added to montmorillonite to model the effect of organic carbon in soil. The aim was to study whether there was a systematic effect in the early-shifting of the absorbance peaks, and if volatilisation losses would occur. Humic acid is one of three generic divisions of soil organic matter, soluble in alkaline solutions and having a relative molecular mass in the range 300-200 000 g mol-1. The acidic functional groups involving oxygen (carboxy , hydroxy , etc.) of humic acid are thought23 to chelate positive ionic species such as Pb2+.As Fig. 4 shows, increasing amounts of humic acid added to the montmorillonite slurry resulted in earlier shifting of the Pb signal. Three possible mechanisms could account for this. Firstly, humic acid could coat the montmorillonite particles and produce active sites, thus aiding the atomisation process. This would be consistent with the active site model developed by Holcombe and co-workers24J5 for reactions on the graphite surface of the ETA tube. Secondly, a humic acid coating on the particles could promote the carbothermal reduction of Pb0.26 Thirdly, humic acid may be acting as an O2 scavenger, thereby reducing the partial 1.0 1 1 4 1800 I 0.8 1 1 I 0 1 2 3 4 5 Timels Fig. 2. Effect of an organic matrix on the Pb absorbance signal: A, an aqueous Pb solution (1 ng Pb); B, 1 ng Pb as lead(I1) cyclohexane butyrate in xylene; and C, the tube-wall temperature.Peak areas are given in parentheses pressure of O2 and promoting the thermal dissociation of Pb0.21727 All three mechanisms may occur to some degree, but the mechanism of O2 scavenging probably dominates. The other two mechanisms require the uniform coating of clay particles by humic acid. This seems unlikely because humic acid is insoluble, existing in the slurry as finely suspended particulate matter. There was no reduction in Pb integrated absorbance with increasing amounts of humic acid. Hence, the volatilisation losses experienced with solutions did not occur. Effect of Magnesium Magnesium salts are well known as matrix modifiers in ETA-AAS.28,29 They can act as charring aids28 or delay atomisation until nearly isothermal conditions exist in the 1 .o 0.8 0, (II 2 0.6 e $ 2 0.4 0.2 0 1 2 3 4 5 "-1 (b' =.2 1600 5 a l- P) E 1400 0 1 2 3 4 5 Time/s Fig. 3. Absorbance signals for two soil slurries: (a) untreated soils; and ( b ) soils with organic carbon removed by heating in a muffle furnace. A , Aqueous Pb solution (1 ng Pb); B, soil 8 (4.7% organic carbon); C, soil 7102 (0.3% organic carbon); and D, the tube-wall temperature 1800 I 1 1500 I 1 I I 1 0 4 8 12 16 20 Humic acid concentration, % Fig. 4. Lead (0.63 ng) in a montmorillonite slurry. Dependence of absorbance Peak time on the concentration of organic carbon (added as humic acid)444 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 electrothermal atomiser .28JO Slavin et ~2.28 suggested that the analyte is embedded or occluded in MgO particles formed from the magnesium salt during the ETA charring stage. The analyte is then released at a higher temperature when MgO vaporises. The effect of adding Mg2+ to the montmorillonite slurry is shown in Fig. 5. The Pb absorbance is delayed further as amounts of Mg2+ of up to 500 ng are added to 0.63 ng of Pb, but adding more Mg2+ has no further effect. There is a small amount (2.7%) of Mg naturally present in montmorillonitelg but it is bound within the matrix by having displaced Al. Therefore, it would have limited surface interaction with adsorbed Pb. Consequently, any delaying effect due to Mg naturally occurring in montmorillonite should be minimal.Saskatchewan soils have typical Mg concentrations in the range 0.1-0.2 pg g-131 (equivalent to 6-12 pg in a slurry aliquot). Therefore, it is clear from Fig. 5 that there is not enough Mg in these soils to have a significant delaying effect on Pb absorbance. Only soils containing mg 8-1 concentra- tions of available Mg would be likely to show an effect. Effect of Phosphate Phosphate has been used frequently as a matrix modifier in ETA-AAS, for example, in the analysis of fish,32 foods33J4 and blood.35J6 Typically, phosphate is added as NH4H2P04 or (NH4)2HP04, and it is particularly effective in delaying the absorbance signal for Pb. Eaton and Holcombe36 reported that NH4H2P04 increased the appearance temperature for Pb in blood by 400 "C as the phosphate concentration was increased (0.001-1% rn/V).Manning and Slavin29 found that NH4H2P04 reduced interference from MgC12 on Pb and allowed a higher charring temperature to be used. It has been suggested37 that the heterogeneous reaction between PbO(s) and P4OlO(g) forms Pb2P207. This pyrophosphate was repor- ted to decompose at 1206 K, so the absorbance signal would be delayed compared with aqueous Pb solutions, which have an appearance temperature of 1050 K. The addition of increasing amounts of (NH4)2HP04 to the montmorillonite slurry had a strong delaying effect on the Pb signal (Fig. 6), i.e., it enhanced the delay already being caused by the clay matrix. In the figure, the amount of phosphate is calculated as pg of Po43-. Typical concentrations of total P in soil vary widely (50-1500 pg g-I),12 and North American soils generally contain >1100 pg g-1 (equivalent to about 0.2 pg Po43- in a slurry aliquot).Therefore, there is unlikely to be enough P present in soils to delay Pb absorbance significantly. Effect of Chloride Chlorides are known to cause interference with Pb determina- tion, probably through losses of volatile PbCl2.3g In order to determine whether soluble chlorides in soil would affect the absorbance signal, the soils and montmorillonite used in this 2100 2000 .-. E a r" 1900 1800 t 0 400 800 1200 1600 2000 Mg2+ addedlng Fig. 5. Lead (0.63 ng) in a montmorillonite slurry. Dependence of absorbance peak time on the amount of added Mg*+ study were tested for soluble chloride by the method outlined in Page et aZ.22 Trace concentrations only (0.03-2 pg g-1) were found.Additions of these concentrations (as NaC1) to montmorillonite had no measurable effect on the Pb absor- bance signal. Hence chloride interference was considered to be negligible. Combined Effects of Humic Acid, Magnesium and Phosphate Of the species examined, only organic carbon and the clay matrix are likely to be present in soil at high enough concentrations to have an effect on the absorbance of Pb. Their observed behaviour suggests that compensating effects are likely; i. e., organic carbon enhances Pb volatilisation, and the clay matrix delays the signal. Hence, it is probable that certain soils with a high organic carbon content might cause Pb to be volatilised prior to the nearly isothermal region.However, the results also suggest that it may be possible to avoid this by further delaying the absorbance signal through the addition of Mg2+ and/or phosphate. The combined effects of these additives were examined by adding a fixed concentra- tion (15%) of humic acid to the montmorillonite and varying the concentrations of Mg2+ and phosphate. Inspection of Fig. 7 shows that when the phosphate concentration is zero, Tp& increases steadily with increasing Mg2+ (unlike the situation in the absence of humic acid), until little further increase is seen after adding 1600-1700 ng Mg2+ per aliquot. However, when the Mg2+ concentration is zero, phosphate has a greater delaying effect on montmorillonite plus humic acid. In conventional solution ETA-AAS, Mg2+ and phosphate 0 4 8 12 16 Phosphate added/pg Fig.6. absorbance peak time on the amount of added phosphate Lead (0.63 ng) in a montmorillonite slurry. Dependence of " 1800 " Fig. 7. Lead (0.63 ng) in a montmorillonite slurry containing 15% humic acid. De endence of absorbance peak time on the combined effects of addecfMgz+ and phosphateJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 445 together have a combined delaying effect. Therefore, it is surprising to see in Fig. 7 that the addition of Mg*+ to montmorillonite already containing 15% humic acid and at least 3-5 pg of phosphate per aliquot does not increase the signal delay. Even more surprisingly, in the presence of larger amounts of phosphate, the addition of Mg2+ reduces the delaying effect of the phosphate.The reasons for this are not clear. However, the results show that either Mg2f or phosphate should be suitable as a matrix modifier when determining Pb in soil by slurry - ETA-AAS. Conclusions In L’vov platform ETA-AAS it is important to volatilise the analyte into a nearly isothermal environment. For this, matrix modifiers are frequently needed. The above results confirm earlier work,’l showing that matrix modifiers are effective in slurry - ETA-AAS, and it is recommended that a matrix modifier should always be used when determining Pb in soil. Otherwise, a systematic error may occur if the soil has a high organic carbon content. It is clear from this work that the direct analysis of solid samples by ETA-AAS would have limited application if matrix modifiers could not be used.However, their effectiveness has yet to be demonstrated when the direct weighing method of solid sample introduction is used. If a condensed phase reaction is involved, then it seems likely that less efficient contact with the analyte will render the matrix modifier less effective. If so, direct weighing would be severely disadvantaged compared with the slurry technique. To date, little work has been done to determine how matrix modifiers work either with solutions or slurries. In Part 2, we will evaluate several matrix modifiers for the determination of Pb in soil slurries by comparing absorbance peak characteris- tics. We thank Drs. R. Karwowska and R. J. St. Arnaud for helpful discussions during this work. The assistance of the Saskatche- wan Soil Testing Laboratory in carrying out the organic carbon determinations is acknowledged.We are grateful to the Natural Sciences and Engineering Research Council of Canada for financial support of this research. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Jackson, K. W., and Newman, A. P., Analyst, 1983, 108,261. Hinds, M. W., Jackson, K. W., and Newman, A. P., Analyst, 1985, 110, 947. Headridge, J. B., Spectrochim. Acta, Part B, 1980, 35, 785. Slavin, W., and Carnrick, G. R., Spectrochim. Acta, Part B, 1984, 39, 271. Fuller, C. W., Analyst, 1976, 101, 961. Ebdon, L., and Pearce, W. C., Analyst, 1982, 107, 942. Lester, J. N., Harrison, R. M., and Perry, R., Sci. Total Environ., 1977, 8, 153. Stephen, S. C., Littlejohn, D., and Ottaway, J. M., Analyst, 1985, 110, 1147.Olayinka, K. O., Haswell, S. J., and Grzeskowiak, R., J. Anal. At. Spectrom., 1986, 1, 297. Karwowska, R., and Jackson, K. W., Spectrochim. Acta, Part B, 1986,41,947. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. Karwowska, R., and Jackson, K. W., J. Anal. At. Spectrom., 1987, 2, 125. Brady, N. C. , “The Nature and Properties of Soil,” Macmillan, New York, 1984. Bolt, G . H., and Bruggenwert, M. G. M., Editors, “Develop- ments in Soil Science 5A, Soil Chemistry A. Basic Elements,” Elsevier, Amsterdam , 1978. Mattigod, S. V., Sposito, G., and Page, A. L., in Dowdy, R. H., Editor, “Chemistry in the Soil Environment,” ASA Special Publication No. 40, American Society of Agronomy and Soil Science Society of America, Madison, WI, 1981.Lighthart, B., Baham, J., and Volk, V. V., J. Environ. Qual., 1983,4, 543. Guy, R. D., Chakrabarti, C. L., and McBain, D. C., Water Res., 1978, 12, 21. Allen, E., and Jackson, K. W., Anal. Chim. Acta, 1987, 192, 355 * Kerr, P. F. , Editor, “Reference Clay Minerals,” American Petroleum Institute Research Report No. 49, Columbia Uni- versity, New York, 1949. Agemain, H., and Chau, A. S. Y . , Analyst, 1976, 101, 761. Hulanicki, A., and Bulska, E., Can. J. Spectrosc., 1984, 29, 148. Sturgeon, R. E., and Berman, S. S., Anal. Chem., 1985, 57, 1268. Page, A. L., Miller, R. H., and Keeney, D. R., Editors, “Methods of Soils Analysis. Part 2. Chemical and Microbio- logical Properties,” Second Edition, American Society of Agronomy and Soil Science Society of America, Madison, WI, 1982. Keeney, D. R., and Wildburg, R. E., in Elliott, L. F., Editor, “Soils for Management of Organic Wastes and Waste Waters,” Soil Science Society of America, American Society of Agronomy and Crop Science Society of America, Madison, WI, 1977. Salmon, S. G., Davis, R. H., and Holcombe, J. A., Anal. Chem., 1981, 53, 324. Salmon, S. G., and Holcombe, J . A., Anal. Chem., 1982, 54, 630. Cambell, W. C., and Ottaway, J. M., Talanta, 1974, 21, 837. Sturgeon, R. E . , Siu, K. W. M., and Berman, S. S . , Spectrochim. Acta, Part B, 1984, 39, 213. Slavin, W., Carnrick, G. R., and Manning, D. C., Anal. Chem., 1982, 54, 621. Manning, D. C., and Slavin, W., Appl. Spectrosc., 1983,37, 1. Manning, D. C., Slavin, W., and Carnrick, G. R. , Spectrochim. Acta, Part B, 1982, 37, 331. Acton, D. F. , and Ellis, J. G., “The Soils of the Saskatoon Map Area 73-B, Saskatchewan,” University of Saskatchewan, Sas- katoon, Canada, 1978. May, T. W., and Brumbaugh, W. G., Anal. Chem., 1982,54, 1032. Koirtyohann, S. R., Kaiser, M. L., and Hinderberger, E. J., J. Assoc. Off. Anal. Chem., 1982, 65, 999. Rains, T. C., Rush, T. A., and Butler, T. S., J. Assoc. Off. Anal. Chem., 1982, 65, 994. Fernandez, F. J., and Hilligoss, D., At. Spectrosc., 1982, 3, 130. Eaton, D. K., and Holcombe, J. A., Anal. Chem., 1983, 55, 946. Czobik, E. J., and Matousek, J. P., Talanta, 1977, 24, 573. L’vov, B., Spectrochim. Acta, Part B, 1978, 33, 153. Paper JA 712 Received January 20th, 1987 Accepted March 30th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200441

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 447 Determination of Selenium in Serum by Electrothermal Atomisation Atomic Absorption Spectrometry with Deuterium-arc Background Correction Barry Sampson Department of Chemical Pathology, Charing Cross Hospital, Fulham Palace Road, London W6 8RF, UK A method for the measurement of selenium in serum by electrothermal atomisation atomic absorption spectrometry is described. A high-intensity hollow-cathode lamp is found to give a similar performance to an electrodeless discharge lamp. A mixture of copper and magnesium nitrates is used as a matrix modifier with platform atomisation. Air is used in the ashing stage to minimise the accumulation of carbon on the platform. The interference from iron and phosphate has been studied and can be overcome by incorporating a delay of 0.7 s in the read period.Calibration was carried out by the standard additions method. The normal amount of selenium present in plasma samples from healthy young adults was found to be 103.5 _+ 21.3 pg 1-1 (1.31 2 0.27 pmol 1-1). Keywords: Selenium determination; electrothermal atomisation; serum; light sources; background correction Selenium determination by graphite furnace atomic absorp- tion spectrometry is complicated by its high volatility. The use of matrix modifiers can increase the pre-treatment tempera- ture from ca. 300 to 1000°C. Several modifiers have been proposed. That most commonly used is nickel nitrate,1-5 but recently Welz et aZ.6 suggested that a mixture of copper and magnesium nitrates may stabilise the different oxidation states of selenium more equally.This may be particularly important in the analysis of biological samples where the selenium is mainly present as SeII, but SeIV is used as a standard. A further complicating factor in selenium assays in biolog- ical samples is the high non-atomic background absorbance making the use of background correction essential in all assays. The commonly used deuterium-arc correction systems may be subject to systematic errors due to the structured background signals from phosphate and iron, which will give rise to negative signals.7-9 Many workers have claimed accurate background correction for serum samples, but not for whole blood or urine, using deuterium-arc background correction.1-3,5 Using a high-resolution peak display, Welz et a/.* have shown anomalous interference from serum when using deuterium-arc correction.This has been confirmed by Neve et al. 10 The present work, using wall and platform atomisation has also given similar results. It is possible to correct the background accurately using Zeeman-effect background cor- rection.Sl1 One possible alternative approach is time resolu- tion of the background and atomic absorption signals. With this approach it is possible to determine selenium in plasma and whole blood using deuterium-arc background correction. A method is described for the assay of selenium in plasma or serum using the copper - magnesium modifier, deuterium-arc background correction and a hollow-cathode lamp as the light source.Experimental Apparatus A Perkin-Elmer 3030 atomic absorption spectrometer was used with deuterium-arc background correction. A Hilger selenium hollow-cathode lamp was used at 15 mA and a slit width (reduced height) of 2.0 nm. An HGA-500 furnace and AS40 autosampler were also used. The furnace was fitted with pyrolytic graphite coated tubes and total pyrolytic graphite platforms. Argon was used as the inert gas and compressed air as the alternative gas. Samples were pipetted with Oxford pipettes into 0.25-ml polystyrene autosampler cups. The pipette tips and auto- sampler cups were rinsed in nitric acid and water before use in the initial studies; they were subsequenfly found not to contribute any contamination and the washing was discon- tinued. Reagents A 1 g 1-1 stock standard solution of selenium, (seleneous acid, BDH Chemicals) was diluted to a working stock solution of 50 mg 1-1 in 1% nitric acid.Working standards containing 25-500 pg 1-1 were prepared fresh daily from this solution. The matrix modifier contained copper(I1) nitrate (1.9 g 1-I), magnesium nitrate hexahydrate (10.9 g 1-1) and 0.1% Triton X-100. Method For the routine determination of selenium in plasma or serum the proposed procedure is as follows. The sample is diluted 1 + 2 in the autosampler cups with water or a selenium standard to give a zero and two additions, respectively. Concentrations of 50 and 100 pg 1-1 are suitable. Using the AS40 pipette, 20 p1 of diluted sample and 10 p1 of modifier are transferred on to the graphite platform. The furnace and autosampler pro- gramme used is given in Table 1.Air or oxygen is used in the ashing stage to minimise the build-up of carbon residues on the platform and to reduce the background absorption at atomisation. The absorbance is measured as peak height, with a read delay of 0.7 s after initiation of the read cycle and an integration period of 3 s. The concentration is obtained from the standard additions data by a linear regression calculation. Accuracy is assured by assay of the reference serum Seronorm batch 105 (Nycomed, Norway) and the inclusion of an inter-batch sample in each batch. Results and Discussion Method Development Lamp selection Electrodeless discharge lamps (EDLs) are normally recom- mended for selenium and arsenic measurements due to their stable and intense light output.l2,13 However, they require an expensive external power supply, which may not be available to all analysts.Some manufacturers are now recommending448 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 Table 1. Electrothermal atomisation furnace and autosampler pro- grammes Time/s Temperature/"C HGA-5OOfurnace- 80 120 550 550 1000 2300 2500 20 AS 40 autosampler- Sample volume Alternative volume Number of replicates Ramp Hold 5 5 30 20 30 10 (internal gas-air) 1 10 20 20 0 4 (internal gas-stop; 1 5 1 5 read) . . . . . . . . . . 20p1 . , . . 10 p1 (matrix modifier) n . . . . . . . . . . L . . .. ,'... . . -.-., . . . . I 0 50.0 - s a n I 0 50.0 1 50.0 0 Ti mels Fig. 1. Output stability of hollow-cathode and electrodeless dis- charge selenium lamps, all measured with 2.0-nm slit width: ( a ) Hilger hollow-cathode lamp; ( b ) Perkin-Elmer hollow-cathode lamp; and (c) Perkin-Elmer electrodeless discharge lamp that EDLs are no longer required for selenium and arsenic, owing to the availability of high-intensity hollow-cathode lamps.Two selenium hollow-cathode lamps, a Perkin-Elmer Inten- sitron lamp and a Hilger lamp, and a Perkin-Elmer EDL were compared for noise levels and output intensity in the 3030 atomic absorption spectrometer. All lamps were operated in accordance with manufacturers recommendations at slit widths of 0.7 and 2.0 nm. The results of this comparison, shown in Fig. 1, demonstrate that the Perkin-Elmer hollow- cathode lamp is notably inferior to the Hilger lamp and the EDL with respect to the level of noise obtained. The instrument energy levels, which are proportional to the gain on the photomultiplier, and hence to lamp emission were: for the Perkin-Elmer hollow-cathode lamp 45, for the Hilger hollow-cathode lamp 67 and for the Perkin-Elmer EDL 73.The Hilger lamp, although showing a higher noise level than the EDL, had a similar intensity and was judged to be 0 3.0 L 2 0.5 a 0 .... \ ......... Y. ......................... . 'I ........... .... ............. 0.95 s 3.0 Ti me/s Fig. 2 Effect of phosphate and iron on atomic absorption signal from the aqueous selenium standard (80 pg 1-1); 20 pl of standard, with matrix modifier added, plus 10 pl of test solution were atomised as in Table 1. A, Corrected atomic absorption signal; B, back round signal.( a ) No added hos hate or iron; (b) 5 mM phosphate; (3 5 mM iron; and ( d ) 5 mM pEospiate plus 5 mM iron adequate for routine selenium assays, although it may lead to a high detection limit due to higher noise levels. The Hilger lamp was used for all work reported here, and has performed satisfactorily in routine use for a considerable time. Similar comparisons were made using a Varian selenium hollow- cathode lamp. This had a comparable performance to the Hilger hollow-cathode lamp. The effect of light output may be less pronounced in an instrument with a simpler optical system than the 3030, and a hollow-cathode lamp with a lower light output may give a satisfactory performance. Amount of modifer The choice of the copper - magnesium nitrate modifier was suggested by the work of Welz et a1.6 and by a desire not to introduce nickel into the furnace housing, as there is a separate need to use the same apparatus for the assay of nickel in biological fluids.Copper is not determined by furnace methods with this apparatus. The effect of varying the copper and magnesium concentration was tested, and the amounts suggested by Welz et al. were found to be adequate. Interferences The effect of phosphate and iron on the signal from an aqueous selenium standard was investigated and the results are shown in Figs. 2 and 3. The addition of phosphate in amounts comparable to those in whole blood (up to 5 mmol 1-l) has a measurable effect on the corrected atomic absorption signal, but by commencing the integration period later in the atomisation cycle the effect of this can be minimised.The same principle applies to high concentrations of iron (up to 10 mmol 1-1). As can be seen in the data presented in Fig. 3, a considerably greater interference occurs with peak-area than peak-height measurements. The effects seen in serum are greater than those seen in the aqueous standards. This may be, in part, due to the lower concentra- tion of selenium in the serum sample used and also to a matrix effect in the serum. It is probable that there is a matrix effectJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 449 6 c 0 1 2 3 4 5 6 7 8 9 1 0 d o L ' ' ' ' ' ' ' ' ' ' (II L I The absorption signal given by plasma when analysed by the recommended method is shown in Fig.4. The assay of whole blood requires a ten-fold dilution to minimise the iron and phosphate interferences, which is on the edge of the reliable detection limit of the method. The concentration of phosphate in urine is considerably higher, up to 70-80 mmol l-1, and the selenium concentration is much less than that in plasma, so that it is difficult to see urine selenium assays ever being practicable by this method. 0 1 2 3 4 5 6 7 8 9 1 0 Concentration of interferent/mM Fig. 3. Effect of increasing concentrations of phosphate and iron on the selenium signal, as in Fig. 2, H = peak-height measurements and A = peak-area measurements: ( a ) aqueous selenium standard, 80 pg I-*; and ( b ) serum with 100 pg 1-1 of selenium 0.4 E..: 1 0: 2.0 Timeis Fig.4. Absorption signal from plasma assayed by the recommended method, with read delay of 0.7 s: solid line, corrected atomic absorption signal; and broken line, background signal Table 2. Recovery of added selenium in plasma Amount added/ Amount found Pg I-' I-' Recovery, YO 0 51.7 50 107.6 111.8 100 149.1 97.5 200 235.1 91.7 Calibration The slope of the standard additions curve in plasma shows a significant difference from that of aqueous standards, necessi- tating the use of serum based standards. In addition, the range of slopes seen with the serum samples can show a wide variation. This may be due to two causes: an incomplete correction of background interference by the deuterium corrector or to changing sensitivity of the graphite tube with age. The latter could be due to a build up of residue, although normally there is no residual carbon on the platform after over 100-200 firings of serum samples in the tube, or to a change of sensitivity due to the effect of oxygen on the tube.Analytical Performance The IUPAC certified plasma Seronorm 10514 is assayed in each batch at least once as a check on accuracy. The target value is 90.7 k 6 kg 1-1, and the result achieved over five batches is 96.9 k 5.9 pg 1-1. A full comparison of this method with hydride generation atomic absorption and molecular fluorescence using the 2,3-diaminonaphthalene derivative has been performed. This is the subject of a separate publicationls and shows a good correlation between all three methods. An evaluation of performance in an independent quality assurance scheme at the University of Surrey also shows an acceptable perfor- mance.Recovery was assessed by spiking 10-ml aliquots taken from a single pool of plasma with selenium. The spiked aliquots were then assayed by the standard additions method. The results are shown in Table 2. Recovery is within the range 9&110%. Within-batch precision was assessed on several occasions and was found to be 2.5% at a level of 130 pg 1-1 of Se. Between-batch precision has been estimated in two ways: the precision as shown by the repeated Seronorm analysis is 6.1%, and using 27 pairs of replicate samples the precision is 9.7% at a mean of 1.3 pmol 1-1. A normal range was established with plasma samples obtained from healthy young adults ( n = 21) to be 103.5 k 21.3 pg 1-1 (1.31 k 0.27 pmol 1-1).different to that of iron and phosphate as is demonstrated by the considerable difference in calibration graphs found for serum and aqueous standards, and in the between-serum differences. The effect of the platform, by delaying the analyte atomisation, is to allow the resolution of the analyte and major interferent peaks to be well within the capabilities of the deuterium system. Although peak-area measurement is com- monly used with platform atomisation, peak height is preferred owing to the increased sensitivity obtained. Peak- height results are not affected by any residual negative base-line effects due to the correction system. The effects of iron and phosphate are additive, and so it is possible only to allow a final concentration of each interferent of less than 1-2 mmol 1 - I .The assay will readily cope with the amounts of phosphate present in plasma at the lower three-fold dilution used in the routine method. Conclusions It is possible to perform satisfactory assays for selenium using a hollow-cathode lamp with a high-intensity light output. Electrodeless discharge lamps are not essential for the determination of selenium by atomic absorption spec- trometry. The use of appropriate temperature programming in the graphite furnace and selection of the optimum atomisation conditions can allow the use of deuterium-arc background correction instead of Zeeman-effect background correction. These factors will all combine to allow laboratories not previously able to assay selenium to do so.Since this work was completed, there have been reports of the value of palladium, with or without added reducing agents, as a matrix modifier for selenium.16-18 Investigations are in progress to compare the use of palladium with the approach reported here.450 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Saeed, K., Thomassen, Y., and Langmyhr, F. J . , Anal. Chim. Acta, 1979, 110, 285. Alfthan, G . , and Kumpulainen, J., Anal. Chim. Acta, 1982, 160, 221. Brown, A. A . , Ottaway, J . M., and Fell, G. S., Anal. Proc., 1982, 19, 321. Dillon, L. J . , Hilderbrand, D. C., and Groon, K. S . , At. Spectrosc., 1982, 3, 5. Oster, O., and Prellwitz, W., Clin. Chirn. Acta, 1982, 124,277. Welz, B., Schlemmer, G., and Vollkopf, U., Spectrochim. Acta, Part B , 1984, 39, 501. Manning, D. C . , At. Absorpt. Newsl., 1978, 17, 107. Welz, B . , Melcher, M., and Schlemmer, G., Fresenius 2. Anal. Chem., 1983, 316,271. Saeed, K., and Thomassen, Y . , Anal. Chim. Acta, 1981, 130, 281. Neve, J . , Chamart, S . , and Molle, L., in Bratter. P., and Schramel, P., Editors, “Trace Element Analytical Chemistry in Medicine and Biology,” Volume 4, Walter de Gruyter, Berlin, 1987, p. 349. 11. 12. 13. 14. 15. 16. 17. 18. Carnrick, G. R., Manning, D. C., and Slavin, W., At. Spectrosc., 1985, 4, 87. Barnett, W. D., Vollmer, J. D., and De Nuzzo, S. M., At. Absorpt. Newsl., 1976, 15, 33. Verlinden, M., Deelstra, H., and Adriaenssens, E., Talanta, 1981, 78, 837. Inhat, M., Wolynetz, M. S., Thomassen, Y . , and Verlinden, M., Pure Appl. Chem., 1986,58, 1063. MacPherson, A. K . , Sampson, B., and Diplock, A. T., to be published. Schlemmer, G., and Welz, B., Spectrochim. Acta, Parr B , 1986, 41, 1157. Ping, L . , Lei, W., Matsumoto, K., and Fuwa, K., Anal. Sci., 1985, 1,257A. Voth-Beach, L. M., and Shrader, D. E . , J . Anal. At. Spectrom., 1987, 2, 45. Paper 5712.5 Received February 25th, I987 Accepted April 30th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200447

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 45 1 Determination of Lead in Atmospheric Aerosols by Electrothermal Atomisation Atomic Absorption -Spectrometry with Direct Introduction of Filters into the Graphite Furnace M. Jose M. P. Moura, M. Teresa S. D. Vasconcelos and Adelio A. S. C. Machado* Chemistry Department, Faculty of Science, P4000 Porto, Portugal An electrothermal atomisation atomic absorption spectrometric (ETA-AAS) procedure for the determination of lead in atmospheric aerosols has been evaluated, in which small discs (2 mm diameter) cut from a filter, on which atmospheric aerosols had been collected by classical equipment, were introduced into the graphite furnace without any previous treatment. The pyrolysis and atomisation temperatures were fixed at 350/750 "C (15/40 s) and 2080/2080 "C (0/10 s).The 283.3-nm line was used for the determination of lead. Under these conditions, the procedure gave the following results for the determination of lead (expressed as mass per filter area): range, up to ca. 25 ng cm-2; detection limit, ca. 0.7 ng cm-2; precision (RSD), 9.3% at 7.3 ng cm-2 and 6.3% at 10.2 ng cm-2. The accuracy of the procedure was confirmed by the acceptable agreement of parallel results obtained by the described procedure and by ETA-AAS determinations on liquid samples prepared by wet digestion of larger pieces of the same filters. Keywords: Atomic absorption spectrometry; electrothermal atomisation; lead determination; atmospheric aerosols; direct introduction of filters Atomic absorption spectrometry with flame (FAAS) or, more recently, with electrothermal atomisation (ETA-AAS), has become a widely used analytical technique for the determina- tion of metals in air samples owing to its sensitivity, low detection limits, accuracy, wide applicability, high selectivity and almost total absence of interferences.Its application has been recently reviewed by Sneddon.1 In most instances, ETA-AAS has been applied after wet digestion of samples collected on filters. Several papers dealing with devices for the direct collection of particulate samples in receptacles specially designed for direct use in electrothermal atomisers have been reported. Deposition in graphite or tantalum cups or rods by filtration through paper2-5 or porous graphite,+" inertial impactation12-14 or electrostatic precipitation1517 have been proposed.The advantages of these procedures are that, as no pre-treatment steps are required, the corresponding errors due, for instance, to losses or contamination are eliminated, and tedious and time-consuming work is saved. However, when inertial impactation or electrostatic precipitation are used, a constant collection efficiency is difficult to achieve and analytical standardisations require special procedures. l 3 ~ 8 Moreover, equipment for direct collection does not appear to be commercially available and its construction requires technical resources not always available to environmentalists. In consequence, from a practical point of view, direct determinations with commercial graphite furnaces on samples obtained with standard sample collection equipment are required.Such determinations require the burning of solid samples in the electrothermal atomiser, a type of procedure which currently deserves much interest in the ETA-AAS field. For the development of a programme that is in progress in this Department,lg the purpose of which is monitoring metals in aerosols of the urban atmosphere of Oporto, a large number of samples collected by classical procedures had to be analysed. As a consequence, procedures to expedite the determinations were considered. The atomisation of solid pieces of filters in the graphite furnace without previous wet digestion was investigated, as no previous studies of such a procedure for this type of sample were found in the literature' (however, Cernik20 has reported the direct determination of lead in blood spotted on to qualitative filter-paper using a carbon rod atomiser).This paper reports the results of an * To whom correspondence should be addressed. evaluation study of an analytical procedure for the direct determination of lead in atmospheric aerosols collected on cellulose filters by a low-volume sampler. The procedure involves the direct introduction of small discs cut from the filter into the graphite tube atomiser. Experimental Apparatus A non-commercial low-volume aerosol sampler (with flow- rates of up to 1.2 m3 cm-2 d-I), kindly loaned by the Lawrence Berkeley Laboratory, University of California, Berkeley, CA, was used to collect aerosols (diameter less than 1.5 pm) on cellulose ester filters (Millipore, Type RAWP 047 00, pore size 1.2 pm).Atomic absorption spectrometric determinations were car- ried out on an IL551 instrument with an IL62927 hollow- cathode lamp and an IL655 graphite furnace. Owing to the problems with the 217.0-nm Pb resonance line,21 the 283.3-nm line was used for AAS measurements. Pyrolytic graphite furnaces (IL43988) and microboats (IL44119) were used for solid samples and IL29652 graphite furnaces for liquid samples. The instrumental conditions used are presented in Table 1. Reagents Spectrosol grade Pb(N03)2 standard solution (BDH Chemi- cals) was used. All other chemicals were of analytical-reagent grade. The water was de-ionised and distilled in a quartz still (Heraeus B1 18 double distillation unit).Procedures For direct measurements, small discs (constant 2 mm diameter) were cut from the filter with a steel circular cutter and placed in the graphite microboats which were introduced into the graphite furnace. The discs were handled with stainless-steel pincers. Three concordant readings were obtained for each filter by repeating this procedure (more readings were obtained for reproducibility studies, see under Evaluation of Analytical Performances). For standardisation, constant volumes (usually 10 pl) of the blank and up to five standard solutions (see below) withJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 ~ a ? ~:>;-~~!.rtrations in the range 50-800 pg were added to bk,? d- filters placed in microboats. For determinations, the ~ a , ~ ~ .; volume of a blank solution was added to the sample d1scs. For comparison, measurements on solutions obtained by wet digestion (with concentrated nitric acid)22 of larger discs of the same filter were performed. These were attacked in calibrated flasks placed in water-baths at 70-80 "C. The destruction af the pieces of filter was complete in 3-4 min. Blanks 'and standards were prepared by a similar procedure from pieces of blank filters. The final nitric acid concentration in all the solutions was 0.42 M, corresponding to the minimum amount of acid required to dissolve the filter pieces. This value was used after a study of the influence of the acidity of the solution on the peak height confirmed23 that a steep signal depression occurred at higher acid concentrations.The following temperature programme was used in these determinations on liquid samples with the graphite furnace: drying, step 1, 70 "C for 5 s, step 2, 110 "C for 45 s; pyrolysis, step 3,350 "C for 15 s, step 4,650 "C for 20 s; atomisation, step 5, 2000 "C for 0 s, step, 6 2000 "C for 10 s. These values are similar to those indicated in the IL graphite furnace instruc- tions24 for other solutions of the same type. Guidance from other available literature23J5.26 was less helpful. Wide varia- tions are found2"25,26 in the pyrolysis and atomisation ranges, probably because optimum ranges depend markedly on the instrumentation used. Results and Discussion Pyrolysis - Atomisation Conditions A preliminary study was performed to optimise the pyrolysis - atomisation conditions.Typical results obtained with solid samples are presented in Fig. 1 (graphs obtained with standards were similar). For pyrolysis times of less than 30 s, smokes were observed during atomisation which resulted from incomplete ashing during the pyrolysis stage. A study of the influence of the 350-750 "C pyrolysis time settings on the signal yielded the following results: 15-30 s, 0.125 A; 25-30 s, 0.134 A; 15-40 s, 0.185 A. As a consequence of these results, the furnace was programmed as indicated in Table 1. With reference to measurements made on liquid samples (see under Procedures) both the temperature and the time of pyrolysis were significantly increased. This was found to be necessary in order to obtain complete ashing in the pyrolysis phase.A flow-rate of nitrogen of about 28 m3 h-1 was maintained Table 1. Atomic absorption spectrometer and graphite furnace conditions for measurements on solid discs 1L5.51 atomic absorption spectrometer- Hollow-cathode lamp current . . 5 mA Wavelength . . . . . . . . 283.3 nm Band width . . . . . . . . 0.5 nm Background correction . . . . Deuterium lamp Measurement mode . . . . . . Peak height Integrationtime . . . . . . 8 s IL655 graphite furnace- Pyrolytic graphite tubes and microboats Liquid standard/blank volume . . 10 p1 Solid disc area (0 = 2 mm) . . 3.14 mm2 Stage -~ Parameter Drying Pyrolysis Atomisat ion Step . . . . 1 2 3 4 5 6 Temperature/ "C . . . . 70 110 350 750 2080 2080 Time/s . . 5 45 15 40 0 10 Flow-rate (N,)/m'h-' 28 28 28 14 14 14 during drying.Higher flow-rates cannot be used with solid samples because they provoke dragging of the disc during drying. Lower flow-rates cause a decrease in the absorbance values, as shown by the following results: 28 m3 h--1,0.056 A; 21 m3 h-1,0.030 A; 17 m3 h--1,0.029 A. Evaluation of Analytical Performance Range and detection limit Calibration graphs were linear up to ca. 800 pg (ca. 25 ng cm-2 if expressed in mass per filter area) of lead. The detection limit was determined by repeated measure- ments on the blanks. Twenty values showed a standard deviation in absorbance units of o = 0.004. The corresponding detection limit (30)27 is 0.012 (when this parameter was obtained from statistical analysis of typical calibration data27 a lower value was obtained). The corresponding value in mass of lead is ca.20 pg or ca. 0.7 ng cm-2. Precision The precision of the procedure was evaluated by repeated measurements on small discs cut from the same filter. In order to test the homogeneity of particle deposition on the filters, 4-5 successive sets of 4-6 discs were taken from different circular crowns of the filter. Each set was considered as a statistical sample for the purpose of an ANOVA calculation.27 As shown by the typical results presented in Table 2, no significative difference was found at the 5% significance level between the different sets of discs. This result indicates that the solid particles are deposited homogeneously on the filter, which allows direct measure- ment, as used in this procedure, and confirms results described in the literature in which larger pieces of filters were analysed after wet digestion.23 In spite of these statistical conclusions, and as shown by the results in Table 2, the dispersion of values in the set of discs taken from the periphery of the deposit on the filter (set A) was found to be larger than for inner (B-D) sets.As a consequence, it is considered to be safer practice not to cut discs from the periphery of the filter for measurements. Relative standard deviations of 93% (n = 14) for a filter with 7.32 ng cm-2 of lead (mass in each disc 230 pg) and 6.3% (n = 27) for a filter with 10.2 ng cm-2 (320 pg in each disc) were obtained. For comparison purposes, a study of the precision of the wet-digestion procedure was carried out. A filter was used which upon attack and dilution to a suitable volume yielded about the same amount of lead (326 pg) as in the second case above (320 pg) when 20 ~1 of the solution were introduced into the graphite furnace.A relative standard deviation of 5.4% was obtained ( n = lo), which a one-tailed F-test27 showed to be significantly smaller than the 6.3% obtained in the direct digestion procedure at the 5% but not at the 1% significance level. In these solution measurements, the signal intensity was I I I // I 1 I I 700 800 1900 2000 2100 Pyrolysis temperaturePC Atomisation temperaturePC Fig. 1. atomisation temperatures Results of the study for optimisation of the pyrolysis and 452JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 453 Table 2.ANOVA study of filter homogeneity. Sets of n discs (320 pg Pb) were cut from the same circular crown of the filter (10.2 ng cm-2 Pb) from the periphery (A) to the centre (E); results are the range and mean values (relative standard deviation in parentheses) in absorbance units ANOVA Set n Range Mean (RSD, %) A . . . . . . 6 0.147-0.188 0.165(9.0) BS ws B . . . . . , 6 0.158-0.187 0.173(6.3) DF . . . . . . 4 22 C . . . . . . 5 0.160-0.181 0.173(4.9) MSSx 104 . . 1.206 1.164 D . . . . . . 5 0.161-0.188 0.177(5.8) F-test . . . . 1.036<2.866 = F4.20,5yA, E . . . . . , 5 0.166-0.183 0.173(3.7) Global . . . . 27 0.147-0.188 0.172(6.3) * BS, between sets; WS, within sets; DF, degrees of freedom; and MSS, mean sum of squares. - m 0.5 7 0.45 5 0.4 2 0.35 0.3 a 5 0.25 -0 ..- 4- .- ti 0.2 F 0 0.15 c 0.1 0.05 - + 4- 0 n a 0 Pb content (wet attackipg m3) Fig. 2. Comparison of the results from direct determination and wet di estion of atmospheric aerosol lead: y = 1.025(+0.096)~ - 0.818(k0.027); r = 0.982; n = 21 found to be similar (more precisely, 20% lower) to that of discs taken from the 7.32 ng cm-2 filter. An F-test27 applied to the standard variations of the absorbance results of these two experiments showed again the better precision of the wet- digestion procedure at the 5% but not at the 1% significance level. Although less precise than the wet-digestion procedure, the proposed direct ashing procedure shows enough precision for the type of determinations that it is intended for. Accuracy The accuracy of the direct determination procedure was evaluated by using as a test procedure the determination of lead by ETA-AAS in solutions obtained from the wet digestion of larger pieces of the same filters.Determinations were carried out in parallel for 21 filters. The corresponding amounts of lead introduced into the furnace were in the range 35-770 pg. The results obtained are presented in Fig. 2. A linear least-squares adjustment of the results yielded the equation [direct determination] = 1.025[wet digestion] -0.018, with correlation coefficient r = 0.982. The standard deviations of the slope and intercept were 0.046 and 0.013, respectively, and the corresponding confidence limits calcu- lated by the t-function27 were k0.096 and k0.027 (at the 5% significance level). In conclusion, this statistical analysis did not show any evidence of relative nor fixed bias in the measured range and the proposed procedure is considered to be acceptable for the purpose of this type of measurement. Examples of results The results shown in Fig.3 are typical of the values obtained for samplings during successive periods of about 1 h during the day. These results show the adequacy of the procedure for the determination of concentration fluctuations at low concentra- tions of metal. The first two filters, which refer to the morning rush period, contained lead concentrations too high to be 1 .o 0.8 m I E 0 0.6 5 2 a C 0 * 0.4 0.2 1 ~ a 1 2 - 10 12 14 16 Time/h Fig. 3. Lead levels in atmospheric aerosols collected at a sampling post on the wall of the Faculty building (height 3 m, 8 m from the street edge) on 14th May, 1986.Air flow-rate: 4.83 x 10-2 m3 h-1 cm-2; sunny weather; 1 and 2 wet attack, 3-8 direct determination procedure determined by direct measurement and the wet-digestion procedure was used instead. Conclusions The procedure proposed in this paper does not require special sampling equipment to allow direct collection into a receptacle suitable for introduction into the graphite furnace, and does not involve time-consuming procedures. The procedure yields reliable results, with a detection limit of ca. 1 ng cm-2. This value is similar to the detection limit obtained by proton activation analysis28 and lower than the corresponding values for atomic emission29 and X-ray fluore~cence3~~3~ procedures, which also allow the direct determination of samples on solid filters. The precision is similar or better than the precision for the other methods.2s-3" One obvious disadvantage of the proposed procedure is that it does not allow the simultaneous determination of several elements.However, the area of filter required for measurement is so small (0.2 cm2 is enough to make 4-5 repeated determinations) that determinations of other elements on the same filter, by the same or other techniques, are possible. The proposed procedure is particularly suitable for determi- nations of very low concentrations of heavy metals in air, o r during short periods (1 h or less) for measurements of concentration fluctuations, both in open or occupational454 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 atmospheres, when only classical sampling techniques are available. Financial support (research projects 24/84 and 9/85) was received from the Rectorate of the University of Oporto. The loan of an aerosol sampler by the Lawrence Berkeley Laboratory, University of California, USA, is gratefully acknowledged. The results in Fig. 3 were obtained by Miss Susana M. R. Sousa. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Sneddon, J . , Talanta, 1983, 30, 631. Matousek, J . P., and Brodie, K. G., Anal. Chem., 1973, 45, 1606. Brodie, K. G., and Matousek, J. P., Anal. Chim. Acta, 2974, 69,200. Noller, B. N., and Bloom, H., Atmos. Environ., 1975,9,505. Noller, B. N., and Bloom, H., Anal. Chem., 1977, 49, 346. Woodriff, R., and Lech, J .F., Anal. Chem., 1972,44, 1323. Lech, J . F., Siemer, D. D., and Woodriff, R., Spectrochim. Acta, Part B, 1973,28, 435. Siemer, D. D., Lech, J. F., and Woodriff, R., Spectrochim. Acta, Part B, 1973, 28, 469. Siemer, D. D., Lech, J. F., and Woodriff, R., Appl. Spectrosc., 1974, 28,68. Siemer, D. D., and Woodriff, R., Spectrochim. Acta, Part B, 1974, 29,269. Lech, J. F., Siemer, D. D., and Woodriff, R., Environ. Sci. Technol., 1974, 8, 841. Roques, Y . , and Mathieu, J . , Analusis, 1973,2,481. Sneddon, J . , Anal. Chem., 1984,56, 1982. Sneddon, J . , Anal. Lett., 1985, 18, 1261. Torsi, G., and Desimoni, E., Anal. Lett., 1979, 12, 1361. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Torsi, G., Desimoni, E . , Palmisano, F., and Sabbatini, L., Anal. Chem., 1981, 53, 1035. Torsi, G., Desimoni, E., Palmisano, F., and Sabbatini, L., Analyst, 1982, 107, 96. Sneddon, J., Znt. Lab., 1986, 18(4), 18. Moura, M. J. M. P., Vasconcelos, M. T. S. D., and Machado, A.A.S.C., “Abstracts of Papers of the 7th European Confer- ence on Environmental Pollution, Funchal, Madeira, 1986,” Paper 21. Cernick, A. A., Brit. J . Ind. Med., 1974, 31, 239. Welz, B., “Atomic Absorption Spectrometry,” Second Edition, VCH, Weinheim, 1985, p. 294. Long, S. J . , Suggs, J. C., and Walling, J. F., J . Air Pollut. Control ASSOC., 1979, 29, 28. Janssens, M., and Dams, R., Anal. Chim. Acta, 1973, 65, 41. Sotera, J. J., Bancroft, M. F., Smith, S. B., Jr., and Corum, T. L . , “Atomic Absorption Methods Manual, Volume 2, Flameless Operations,” IL, Wilmington, MA, 1981. Omang, S. H., Anal. Chim. Acta., 1971, 55, 439. Geladi, P., and Adams, F., Anal. Chim. Acta, 1978, 96, 229. Miller, J. C., and Miller, J. N., “Statistics for Analytical Chemistry,” Ellis Horwood, Chichester, 1984, pp. 57-59, 67-74 and 96-100. Desaedeleer, G . , Ronneau, C., and Apers, D., Anal. Chem., 1976, 48, 572. Scott, D. R . , Hemphill, D. C., Holboke, L. E., Long, S. J . , Loseke, W. A., Pranger, L. J., andThompson, R. J . , Environ. Sci. Technol., 1976, 10,877. van Espen, P., and Adams, F., Anal. Chim. Acta, 1974,7561. Harrison, R. M., and Laxen, D. P. H., “Lead Pollution, Causes and Control,” Chapman and Hall, London, 1981, p. 159. Paper J6112.5 Received December 22nd, I986 Accepted March 30th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200451

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 455 Determination of Cadmium, Copper, Manganese and Rubidium in Plastic Materials by Graphite Furnace Atomic Absorption Spectrometry using Solid Sampling Uwe Vollkopf, Raimund Lehmann and Dietmar Weber Bodenseewerk Perkin-Elmer 8t Co. GmbH, Postfach I 120, 0-7770 Uberlingen, FRG The direct analysis of solid samples using (GFAAS) graphite furnace atomic absorption spectrometry (GFAAS) has considerable potential as a method for the determination of trace metals when only small amounts of sample are available, when the highest accurracy is not necessarily required, or when sample dissolution would be difficult and/or time consuming. Whereas the complete dissolution of many plastics is very difficult and often time consuming, the analysis by solid sampling offers the potential for avoiding sample dissolution.Using the cup-in-tube technique, Cd, Cu, Mn and Rb have been determined in solid plastic samples using Zeeman-effect GFAAS. The analytical conditions used permitted the determinations to be performed using aqueous reference solutions. Acceptable agreement of the results with reference values was achieved. Keywords: Graphite furnace atomic absorption spectrometry; solid sampling; plastics; cadmium, copper, manganese and rubidium determination In the analysis of plastic materials the determination of metals, especially Cd, is of great importance, as they can be present in such materials as catalysts and stabilising agents. Plastic materials are used in a large variety of products today, such as furniture, cloth, toys and automobiles.It is important to know their metal contents for several reasons, including health and environmental protection. Almost all procedures described in the literaturel-4 for the determination of metals in plastics are based on sample decomposition prior to analysis by AAS. All the procedures reported for the decomposition of plastic samples are time consuming and frequently unreliable due to incomplete dissolution of the solid material, contamination or analyte losses. For this reason many analysts who are dealing with the analysis of plastic materials are interested in the sold-sampling technique. The direct analysis of solid samples using GFAAS has considerable potential as a method for the determination of trace metals when only small amounts of sample are available, when the analyst is interested in the distribution of an analyte in certain sample materials, e .g . , gun shot5 residues or bones, when the highest accuracy is not required or when sample dissolution is difficult and time consuming. However, the time saving benefit of solid sampling is very dependent upon the type of solid material and the analyte of interest. With solid- sampling techniques, additional sample preparation may be required to ensure adequate sample homogeneity. Method development, including the optimisation of furnace para- meters and selection of matrix modifiers may be more time consuming than if aqueous samples were used. It is frequently necessary, owing to the extreme sensitivity of the graphite furnace, to use alternative, less sensitive lines, and/or internal gas flows during atomisation to reduce the sensitivity.Most of the early work with solid sampling in graphite furnaces was performed by placing solid samples on the inner tube wall of the graphite furnace.69 Owing to the less sophisticated furnace technology at that time, direct calibra- tion using simple aqueous reference solutions was very often impossible. The different volatilities of the analyte in the reference solution and in the solid sample, spectral inter- ferences and matrix effects leading to signal depression frequently caused erroneous results. With modern furnace technology and applying the concept of the stabilised tem- perature platform furnace (STPF) which was developed by Slavin ef al.") in 1981, the direct analysis of solid samples became more attractive because aqueous reference solutions can frequently be used for instrument calibration. Some workers, however, used only a few of the concept parameters, mainly the platform, for their solid-sampling work.This is most probably the reason that they were able to eliminate matrix interferences completely. 11-13 Grobenski and co-work- ers have shown that for the determination of trace metals in biological materials14 and for Pb in rock samples15 the direct analysis of solid samples against acidified reference solutions was possible if optimised conditions were used in a stabilised temperature platform furnace. In two recent publication~16~17 we described a new solid sampling device (cup-in-tube technique) which allows the convenient introduction of solid samples into the graphite tube of a Zeeman-effect graphite furnace.The solid-sampling cup offers the same advantages as a platform in that it delays the analyte atomisation until the tube wall and the gas atmosphere inside the tube have almost reached a steady-state temperature. Applying this technique, the direct determina- tion of trace elements in soil and hay samples was possible. Carnrick et al. 18 successfully applied the same technique to the determination of Cu, Cr and Pb in plastic film, flexible PVC and bovine liver. In this paper the determination of Cd, Cu, Mn and Rb in different plastic materials applying the cup-in-tube technique and graphite furnace AAS with Zeeman-effect background correction is described.Experimental Instrumentation A Perkin-Elmer Zeeman/3030 system with an HGA-600 graphite furnace and an AS-60 autosampler was used. A Perkin-Elmer AD-6 autobalance was used for weighing the solid samples. Special pyrolytically coated graphite tubes, which accommodate the solid-sampling cups (Perkin-Elmer B016-2706), and pyrolytically coated solid sampling cups (Perkin-Elmer B016-2704) were used throughout this work. The solid sampling cups were inserted into and removed from the furnace tube using a special plastic tool (Perkin-Elmer B014-4618). The design and handling of this device has been described previously. 16-18 The flame AAS determinations were performed using a Perkin-Elmer Model 3030B spec- trometer.456 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 Table 1. Graphite furnace parameters for the determination of Cu, Mn and Rb in polyester, nylon and perlon. Matrix modifier: 5% H2S04 + 3% HNO, (10 PI); RT = ramp time, HT = hold time Cu (244.2 nm) Mn (403.1 nm) Rb (780.0 nm) Tem- perature/ Step "C 1 . . . . 100 2 . . . . 200 3 . . . . 320 4 . . . . 550 5 . . . . 1100 6 . . . . 2300 7 . , , . 2650 RT/ 2 15 25 25 20 0 1 S HT/ 20 10 25 20 20 3 5 S Internal flow/ ml min-1 300 300 300 300 300 50 300 Tem- perature/ "C 100 200 320 550 1000 1900 2650 RT/ 2 15 25 25 20 0 1 S HT/ 20 10 25 20 20 5 5 S Internal flow/ ml min-1 300 300 300 300 300 200 300 Tem- perature/ "C 100 200 320 550 800 1900 2650 RTI 2 15 25 25 20 0 1 s HTI 20 10 25 20 20 5 5 S Internal flow/ ml min-1 300 300 300 300 300 200 300 Table 2.Graphite furnace parameters for the determination of Cd in polymers. Matrix modifier: 5% H2S04 + 3% HN03 + 2% (NH&HP04 (10 PI); RT = ramp time, HT = hold time Tem- Internal perature/ RT/ HT/ flow/ 1 . . . . . . . . . . 100 2 20 300 2 . . . . . . . . . . 200 20 20 300 3 . . . . . . . . . . 400 25 25 300 4 . . . . . . . . . . 700 25 30 300 5 . . . . . . . . . . 1900 0 3 150 6 . . . . . . . . . . 2650 1 5 300 Step "C S s mlmin-1 Samples and Reagents Polyester, perlon and nylon samples, pre-analysed by the supplier (X-ray fluorescence) were analysed for Cu, Mn and Rb. These samples were additionally analysed by flame AAS after an appropriate sample dissolution had been performed. We also participated in an inter-laboratory study on plastic analysis arranged by the German automobile industry in which six different polymer samples (PVC and polypropylene) were analysed for Ca.Reference solutions were prepared from stock standard concentrates (Merck, Darmstadt, FRG) by appropriate dilution using 0.2% mlV HN03. For matrix modification and as a reagent for improving ashing of the plastic samples in the furnace, a mixture of 5% mlV H2S04 and 3% mlVHN03 was used. For dissolution of the polyester, perlon and nylon samples we used a Buchi Digester (Buchi Digester 445, Fa. Buchi, CH-9230 Flawil, Switzerland). Sample Dissolution Procedures Although the polyester, perlon and nylon samples supplied to us were pre-analysed, we were interested in comparing a flame AAS method with the solid-sampling technique.For the decomposition of perlon and nylon approximately 1 g of the samples, 1 ml H2SO4 (concentrated) and 3 ml HN03 were placed in the quartz flasks of the digester. The temperature of the salt heating bath was set to 300°C. The system was programmed in such a way that the quartz flasks were automatically inserted into the salt bath for 2 s, then removed from the bath for 20 s (cooling) before re-insertion. This sequence was repeated for a period of 15 min. After cooling the sample solutions to room temperature, they were quanti- tatively transferred into 100-ml calibrated flasks and diluted to volume with de-ionised water. The very short insertion times into the hot salt bath were necessary due to the strong chemical reactions of the acid mixture with the samples.The polyester sample was also decomposed using the digester; however, a different procedure was used. A "percolator" (an extraction device) was placed into the upper part of each of the quartz flasks, after 5 ml HN03 (concentrated) had been pipetted into the flasks. The sample was deposited on to the bottom of the percolator. The digestor programme was the same as described above, but was repeated for a period of 20 min. During this period the solid samples were completely extracted into the acid solution. After cooling the sample solutions to room temperature, they were quantitatively transferred into 100-ml calibrated flasks and diluted to volume with de-ionised water. Analytical Conditions The determinations by flame AAS were performed using the conditions recommended by the manufacturer.19 The mixing chamber of the burner system was equipped with an adjust- able impact bead which, for a number of analytes, provides significantly increased sensitivity compared with operation with a flow spoiler. The optimised furnace conditions used for the solid-sam- pling technique are listed in Tables 1 and 2. For the determination of Cd, Cu and Mn the graphite furnace sensitivity was much too high. For this reason alternative, less sensitive, resonance lines were used. Further reduction of the furnace sensitivity was achieved by atomising the samples with an internal flow of argon passing through the furnace. To overcome the problem of sample boiling and sample losses during the pre-treatment steps in the furnace, the tube temperature was carefully ramped during the first four temperature - time programme steps.It was found that when 10 pl of a mixed acid (3% HN03 + 5% H2SO4) were added to the top of each sample, the ashing of the plastic material in the furnace could be significantly improved and background absorption was reduced to a minimum. For the determination of Cd, 2% (NH4)2HP04 was added to the acid mixture, allowing a maximum thermal pre-treatment temperature of 700°C without any risk of loosing the analyte prior to atomisation. Sub-samples (0.2-1.5 mg) were prepared by cutting small pieces from the main sample using a scalpel. This procedure did not create any contamination problems, as the concentrations in the samples were very high. Results and Discussion Analysis of Nylon, Perlon and Polyester Applying the optimised graphite furnace parameters shown in Table 1, individual solid samples with masses ranging from ca.0.2 to 1.6 mg were analysed. The spectrometer was calibrated using aqueous standard solutions. All concentration calcu- lations were based on integrated absorbance (peak area) measurements. From every sample type ten sub-samples were prepared for the determination of each element. Although the atomisation signals for the solid samples appeared slightly delayed relative to the signals produced by the aqueous reference solutions, as shown in Fig. 1 for Cu, directJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 457 . AA 0.3 $ BG 0.1 -e 2 m 0 - AA 0.3 BG 0.1 Table 3. Results for the determination of Cu, Mn and Rb in polymer samples ( n = 10).All results are in pg g-1 RSD, Reference Sample Element Result SD YO FAAS value Perlon . . Cu 31 4 12.9 33 30 Mn 22 1.3 5.9 25 20 Rb 1.5 0.16 10.6 1.6 1.5 Nylon . . Cu 66 2 3.0 67 60 Mn 9.7 0.8 8.2 9.7 10 Rb 3.0 0.09 3.0 2.9 3 Polyester . . Mn 57 2 3.5 55 55 -(b) . . . . . . . . . . . . . . . . . . ' - (C) AA 0.4 BG 0.1 I ( a ) Reference solution (100 ng Cu) ........ 0 Perlon l(1.61 mg) rn I ...... . . . . . . n 0- * AA 0.4 (c) Nylon 1 (0.97 mg) BG 0.1 t 0 3.0 Ti m eis Fig. 1. Atomisation signals for Cu in an acidified reference solution and two solid samples: solid line, analyte absorbance; and broken line, background attenuation measurements against these reference solutions were possible.In general there was good agreement between the results found by solid sampling, those found by flame AAS and the reference values. The results are listed in Table 3. Precision values of 3.1-12% were obtained. Determination of Cadmium in Plastic Samples According to new laws which have been passed recently, or which will be passed within the near future by the European Community, the maximum concentration of some toxic heavy metals in plastic material used in car manufacturing, predomi- nantly Cd (75 mg kg-I), will be limited for environmental protection reasons. This means that in the future all automo- bile manufacturers will have to guarantee that the concentra- tion of Cd in the materials used is below the legal limit. Because analysis by flame AAS after sample dissolution is time consuming and not always reliable, there is great interest in a fast, reliable method based on solid sampling.In order to establish such a method for quality control analysis, an inter-laboratory trial was organised where six different poly- mer samples were distributed to a number of different laboratories for analysis. The samples were analysed using different analytical techniques including FAAS, ICP-AES (applying six different dissolution techniques), X-ray fluores- cence and graphite furnace AAS using either solid sampling or dissolved samples. The analytical conditions used for our investigations are listed in Table 2. Owing to the high thermal pre-treatment temperature of 700 "C only low background absorption (less than A = 0.1) was observed in all measurements.Use of the recommended atomisation temperature of 1600 OC20 resulted in multiple atomisation signals for Cd. The temperature had therefore to be optimised for the solid-sampling technique. Fig. 2 shows high resolution atomisation signals for Cd in one of the samples using 1800 and 1900 "C, which was found to be Conclusion It has been shown that the direct analysis of solid plastic samples using GFAAS has considerable potential as a method for quality control purposes. The technique is relatively fast, as time-consuming sample dissolution procedures can be avoided. While sample dissolution using the technique des- cribed in this paper takes ca. 25 min the determination by solid sampling takes ca. 15 min (five replicates per sample).Although the sample decomposition may be automated, the high investment cost for such a decomposition system must be considered. Furthermore, the risk of contamination from the reagents used is likely with any sample decomposition procedure. A further advantage of solid sampling is that it requires only a small amount of sample. The application of STPF conditions in this study resulted in the elimination, or at least in a drastic reduction, of matrix interferences enabling measurements to be made directly against acidified reference solutions when integrated absorbance data were used for signal evaluation rather than peak height. By adding acids for matrix modification, background absorption could be mini- mised.458 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 Table 4. Results for the determination of Cd in two polymer samples Sample 1 Sample 5 Cd concen- Cd concen- Sample trationl Sample tration/ mass/mg g-l masslmg I % g-l 0.85 0.92 1 .OO 0.46 0.69 0.52 0.68 0.49 0.29 0.30 Mean SD . . RSD , O/o 52.8 47.9 53.7 54.1 49.1 50.0 47.2 53.7 47.2 50.0 . . 50.6 . . 2.8 . . 5.5 0.51 0.23 0.35 0.86 0.55 0.37 0.27 0.42 0.36 0.35 Mean SD . . RSD, Yo 994 1013 1051 902 976 1008 893 1000 978 879 . . 969 . . 58 . . 6.0 Table 5. Comparison of the results for the determination of Cd in polymer samples using solid sampling with the results of the inter-laboratory study. All values are in pg g-1 Inter-laboratory study Sample Found (mean f SD, n = 10) 1 . . . . . . 50.6k2.8 56.5 f 3.2 2 . . . . . . 92.6k 12.4 111.5 f 8.1 3 .. . . . . 121.3k2.9 135.8 2 7.4 4 . . . . . . 66.4If: 1.7 69.3 f 3.8 5 . . . . . . 969 k 58 973 k 30 6 . . . . . . 2770k 142 2817 k 101 I I 0.5 1 .o Sample mass/mg Fig. 3. Linear regression graphs for Cd in polymer samples 1-4. A, Sample 1, r = 0.9874; B, sample 2, r = 0.8926; C , sample 3, r = 0.9882; and D sample 4, r = 0.9967 The gain in sensitivity obtained by the solid-sampling technique was not required. In fact sensitivity was reduced by selection of secondary, less sensitive resonance lines and an internal gas flow during atomisation. Owing to the high analyte concentrations in the samples under study no contami- nation problems were observed. The precision is typically better than 10% (relative standard deviation) which demonstrates that the analytes in the samples investigated are distributed homogeneously.1000 0) t 5 3 0 .c 500 E e, 1. (D C w - Q 0 I 0.5 1 .o Sample mass/mg Fig. 4. A, sample 5, r = 0.9886; and B, sample 6, r = 0.9894 Linear regression graphs for Cd in polymer samples 5 and 6: Although the results presented in this paper do demonstrate that by applying the cup-in-tube technique some metals can be determined successfully in a number of plastic samples, it should not be concluded that all metals which can be determined by GFAAS can be measured equally well in any type of plastic material. The applicability of the solid-sampling technique to other analytes or sample matrices should first be verified by comparison with established procedures. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.16. 17. 18. 19. 20. References Oliver, M., Fresenius Z . Anal. Chem., 1969, 248, 145. Oliver, M., At. Absorpt. Newsl., 1971, 10, 12. Girgis-Takla, P . , and Chroneos, I . , Analyst, 1978, 103, 122. Anwar, J . , and Marr, I. L., Talanta, 1982, 29, 869. Vollkopf, U., unpublished data. Kerber, J. D., At. Absorpt. Newsl., 1971, 10, 104. Welz, B., Fortschr. Mineral., 1972, 50, 106. Barnett, W. B., and Kahn, H. L., Clin. Chem., 1972, 18,923. Langmyhr, F. J., Analyst, 1979, 104, 993. Slavin, W., Carnrick, G. R., and Manning, D. C., At. Spectrosc., 1981, 2, 137. Chakrabarti, C. L., and Li, W. C. , Spectrochim. Acta, Part B , 1980,35, 93. Chakrabarti, C. L., Wan, C. C., and Li., W. C., Spectrochim. Acta, Part B , 1980,35, 547. Frech, W., Lundberg, E., and Barbooti, M. M., Anal. Chim. Acta, 1981, 131, 45. Grobenski, Z., Lehmann, R., Tamm, R., and Welz, B., Mikrochim. Acta, 1982, I, 115. Grobenski, Z., and Lehmann, R., Ar. Spectrosc., 1983,4, 111. Vollkopf, U . , Grobenski, Z . , Tamm, R., and Welz, B., Analyst, 1985, 110, 573. Vollkopf, Y., and Grobenski, Z . , “Fortschritte in der atorn- spektrometrischen Spurenanalytik,” Band 2, VCH Verlagsges- chellschaft mbH, Weinheim, 1986. Carnrick, G. R., Lumas, B. K . , and Barnett, W. B., J. Anal. At. Spectrom., 1986, 1,443. “Analytical Methods for Flame AAS,” Perkin-Elmer, Nor- walk, CT, 1982. “Techniques in Graphite Furnace Atomic Absorption Spectro- photometry,” Perkin-Elmer, Ridgefield, CT, April 1985. Paper J6II2 7 Received December 30th, 1986 Accepted April 22nd, I987

ISSN:0267-9477    

DOI:10.1039/JA9870200455

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 459 Electrothermal Atomisation Atomic Absorption Spectrometric Determination of Silver, Beryllium, Calcium, Iron, Lead and Tin in Uranium without Preliminary Separation Neelam Goyal, P. J. Purohit, A. R. Dhobale, B. M. Patel, A. G. Page and M. D. Sastry Radiochemistry Division, Bhabha Atomic Research Center, Tromba y, Bombay 400 085, India Electrothermal atomisation atomic absorption spectrometric methods have been developed for the direct determination of Ag, Be, Ca, Fe, Pb and Sn in uranyl nitrate solution. Using 5-p1 sample aliquots containing 100 pg of uranium, the detection limits for the analytesvary between 0.1 and 25 p.p.m. (on a uranium mass basis). The precision of determinations at an intermediate concentration was calculated to be better than 9% RSD.Interference studies carried out include the effect of the build-up of uranium in the atomiser and the effect of concomitant and major impurities on the analyte absorbance. A measure of the high accuracy of the determinations can be inferred from the analysis of a set of NBL reference standards. Keywords: Electrothermal atomisation; atomic absorption spectrometry; uranium analysis; interference studies Atomic absorption spectrometry utilising electrothermal atomisation devices (ETA-AAS) has proved to be an efficient method for the determination of a number of trace elements in different types of samples. The high sensitivity of ETA-AAS makes it suitable for determining trace constituents in the presence of milligram amounts of the matrix material.This is of significant practical advantage when dealing with radioac- tive materials. The advantages that ETA-AAS offers for radioactive samples have not yet been fully exploited. A number of workers'.* have, however, utilised ETA-AAS for the determination of impurities in radioactive samples after their chemical separation and pre-concentration. We have developed methods for determining impurities in uranium without prior chemical separation of the major matrix and have also investigated the inter-element and matrix effects. s5 This paper is the culmination of our work on uranium and describes the methods developed for the direct determination of Ag, Be, Ca, Fe, Pb and Sn in uranium solution at nanogram and sub-nanogram levels.In this work the effects of the uranium matrix and its residual build-up in the tube and the presence of other metallic elements on the precision and accuracy of the determinations have also been investigated. Experimental Apparatus A Varian Techtron atomic absorption spectrometer (Model AA-6) equipped with a carbon rod atomiser (CRA-63) and BC-6 background corrector was used. The instrumental details are the same as those reported previously.s5 Varian Techtron single element hollow-cathode lamps were used for the determination of Ag, Be, Fe, Pb and Sn while a bielement Ca - Mg lamp was used for the determination of Ca. All lamps were operated at the prescribed currents except for Ca, where the current was kept at 15 mA. The spectral band widths were 1 nm for Ca, 0.5 nm for Ag, Be, Pb and Sn, and 0.2 nm for Fe.An Eppendorff 5-p1 pipette with disposable PTFE tips was used for dispensing the sample solutions. Reagents Water, doubly distilled in quartz apparatus, and electronic grade nitric acid were used for the preparation of standards and samples. Stock solutions (2 mg ml-1) of each of the analytes were prepared by dissolving appropriate amounts of Specpure chemicals (Johnson Matthey, London, UK) in 3 M nitric acid. Working standards were prepared by appropriate dilution of the stock standard with 0.1 M nitric acid. A stock solution of uranium (200 mg ml-1) was prepared by dissolving high-purity uranium in concentrated nitric acid followed by dilution with 0.1 M nitric acid. All solutions were stored in poly(viny1 chloride) containers.For the study of the direct determination of Ag, Be, Ca, Fe, Pb and Sn in the uranium matrix, another set of standard solutions was prepared each containing 20 mg ml-1 of uranium and varying concentrations of the analytes in the range of 0-2 pg ml-1. These standards also contained varying concentra- tions of 18 other metallic elements of interest in the fuel specification analysis, i.e., Al, B, Bi, Cd, Co, Cu, Cr, Li, Mg, Mn, Mo, Na, Ni, Si, Sn, Ti, V and Zn in the concentration range 0-2 pg ml-1 (the zero concentration of trace constitu- ents signifying the blank). Procedure The experimental parameters such as temperature, time duration for the drying, ashing and atomisation stages were optimised using an intermediate analyte concentration stan- dard for all the elements with and without uranium.The entire series of standards were then studied for each analyte element to determine the analytical response. The working graphs obtained in this way were then used for the determination of the analyte concentration in the reference samples. The reproducibility of the method was established by the repetitive analysis of an intermediate analyte concentration standard within the analytical range. Results and Discussion Effect of Uranium Matrix and its Build-up The effect of the uranium matrix on the atomisation behaviour of the analytes was examined to ascertain whether any matrix interferences occurred. Measurements of atomic absorption signals for each of the analytes were carried out over a wide Table 1.Characteristic concentrations (0.0044 A) for the analytes with and without uranium Characteristic concentration/ I.18 cL1-I Wavelength/ Element nm Ag . . . . . . 328.1 Be . . . . . . 234.9 Ca . . . . . . 239.8 Fe . . . . . . 371.9 Pb . . . . . . 283.3 Sn . . . . . . 286.3 Aqueous Uranium solution solution 0.17 0.52 0.15 0.22 100.0 250.0 3 .O 17 .O 1 .o 2.4 3.0 4.50460 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL. 2 range of concentrations using the standard solutions with and without the uranium. The characteristic concentrations6 for these analytes with and without uranium are shown in Table 1. It can be seen from these results that for all the analytes the sensitivity is poorer in the presence of uranium compared with that in its absence for all the analytes.It is essential therefore, to match the uranium contents of the sample solutions in order to analyse uranium samples directly. The non-specific absor- bance due to uranium was monitored using a hydrogen continuum lamp and was found to be insignificant. The effect of uranium build-up in the tube on the analyte signal was also investigated. Whereas for Ag and Be, absorbance signals were found to be essentially unaffected, with the repetitive use of the tube over 100 firings, the analyte signals for Ca, Fe, Sn and Pb were found to reduce by about 30% after 30-40 atomisation cycles. Hence these elements were determined with prior "conditioning" of the tube for the uranium matrix. Conditioning of the tube essentially involves the cumulative firing of 4 rng of uranium in the tube over a number of atomisation cycles.When the concentration of uranium is kept at 20 mg ml-1, as for most of the aliquots used, this amounted to ca. 40 atomisation cycles. However, when a higher concentration of uranium (200 mg ml-1) was used, the same effect could be observed after just four atomisation cycles. Therefore, the conditioning of the tube for the determination of the specified analytes in uranium was carried out by four atomisation cycles, each one involving the loading of 5 pl of a 200 mg ml-1 uranium solution. Effect of Impurities The analytes studied were found to be completely free from any significant inter-element effects when they were all present at trace concentration levels. However, as the maximum permissible concentrations of Ca, Fe, Si, W and Zn in reactor-grade uranium are generally higher than those studied here, typically of the order of 500-1000 p.p,m., the effect of 1000 p.p.m.of each of these five elements on the analytes absorbances was studied using a synthetic sample prepared with the following concentrations: Ag and Be, 1 Table 2. Interference studies on the determination of the analytes Absorbance measurement in presence of 1000 p.p.m. of major contaminants in fuel Analyte Element absorbance Ag . . . . 0.35 Be . . . . 0.54 Ca . . . . 0.15 Fe . . . . 0.26 Pb . . . . 0.45 Sn . . . . 0.30 * Sum of Ca, Fe, Si, Wand Zn together, each at 1000 p.p.m. Ca Fe Si W Zn Zimp.* 0.34 0.30 0.30 0.30 0.28 0.31 0.52 0.48 0.48 0.56 0.48 0.53 - 0.04 0.17 - 0.15 0.04 0.24 - 0.16 0.27 0.14 0.21 0.38 0.45 0.46 0.39 0.41 0.38 0.26 0.28 0.31 0.29 0.25 0.25 Table 3.Analytical results and experimental parameters for uranium matrix Carbon rod atomiser setting Element and line/ nm Ag, 328.1 Be, 234.9 Ca, 239.8 Fe, 371.9 Pb, 283.3 Sn, 286.3 Ash Atomise Temperature/ "C 250 250 250 250 250 250 Time/ 50 50 50 50 50 50 S Temperature/ "C 900 900 900 900 900 900 Time/ 40 40 40 40 40 40 S Temperature/ "C 2250 2700 2550 2550 2400 2400 Time/ 3 3 4 3 3 3 S * Based on 100 pg of uranium in 5 yl of solution. Linear analytical range,* p.p.m. 0.1-4.0 0.1-2.0 25-500 2-100 1-20 1-40 Limit of quantitative determinatiodg Present Reference 2 x 10-1" 2 x 10-1" 2 x 10-10 2 x lo-' work 2 10 x 10-12 10 x 10-12 2.5 x 10-9 - 2.0 x 10-1" 1.0 x 10-1" - 1.0 x 10-1' ~~~ ~~~ ~ Table 4.Analysis of NBL reference standards (p.p.m.) for 10 replicate determinations Sample No. 98-2 98-3 98-4 98-5 Present Certified Present Certified Present Certified Present Certified Element work value work value work value work value f 0 . 1 f0.2 k 0.1 f O . 1 f0.03 k0.06 fO.01 k0.03 Be . . . . . . . . 7.0 9.0 3.7 4.2 2.1 2.2 0.9 0.9 k0.6 k0.06 50.3 20.03 k0.04 k0.03 k0.02 f0.03 Ag . . . . . . . . 2.3 2.3 1.1 1.2 0.4 0.4 0.12 0.2 Ca . . . . . . . . 35.0 f0.3 Fe . . . . . . . . 200 516 Pb . . . . . . . . 16.0 k0.2 Sn . . . . . . . . 13.5 k0. 1 37.1 f l . O 210.2 k3.9 19.0 f1.2 17.6 k0.7 <25 18.8 k0.7 125 109.8 25.5 f 3 . 0 8.8 10.7 f0.8 f0.9 8.5 9.6 20.8 f1.5 <25 8.3 f 0 . 3 60.0 58.9 k3.3 k1.9 4.3 5.2 k0.4 f 0 . 9 3.7 4.7 k0.4 k0.4 <25 3.6 k0.3 30.0 32.9 k2.7 k1.9 2.3 3.0 k0.2 f0.5 2.0 1.9 k0.2 f0.7JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, AUGUST 1987, VOL.2 461 p.p.m.; Fe and Sn, 10 p.p.m,; Pb, 12.5 p.p.m.; and Ca, 375 p.p.m. These concentrations were chosen in order to achieve significant absorbance signals. The results of ten repetitive measurements from these studies are shown in Table 2. The change in the mean analyte absorbance caused by the presence of each of the interferents as shown in Table 2 indicates no significant effect due to the presence of 1000 p.p.m. of either Ca or W. However, Zn and Si at these concentration levels do affect the absorbance signal for Fe significantly and the presence of 1000 p.p.m. of Fe drastically suppresses the absorbance signal for Ca. The presence of Si or Zn at 1000 p.p.m.reduced the absorbance due to 10 p.p.m. of Fe by ca. 50%. When they are present together along with Ca and W each at the 1000 p.p.m. level, the reduction in the intensity of Fe is ca. 16% (Table 2). Therefore, the effect of Fe on the absorbance of Ca is more pronounced. Our results suggest that for up to 400 p.p.m. of Fe there is no substantial reduction in the absorbance of 350 p.p.m. of Ca. Beyond 400 p.p.m. of Fe there is a gradual reduction in the absorbance of Ca. In view of this, prior knowledge of the concentrations of Zn and Si is essential for the determination of Fe while estimates of the Fe content are essential for the determination of Ca in fuel samples. Analytical Results The analytical range and the lowest absolute amounts determined for each of the analytes in the uranium matrix are given in Table 3.The limits of the quantitative determinations in uranium obtained in this work are lower by a factor of 20 for Ag and Be, and by a factor of 200 for Sn than those reported by Buffereau and Robichet2 while for Fe the limit of determination is comparable. Similar limits for Ca and Pb have not yet been reported in the literature. The precision of the determinations for the six analytes at different concentrations was calculated from ten peak-absor- bance measurements and found to be 9% RSD or better. The precision obtained from ten replicate measurements are as follows: lower concentration end of the working curve, &lo% RSD; middle of the working range, 3-5% RSD; and the high concentration limit of the working range is kept well below the non-linear region and the precision is ca.5% RSD. Using the standardised procedure, the six analytes were determined in a set of four U308 reference standards obtained from the New Brunswick Laboratory (NBL Atomic Energy Commission, New Brunswick, NJ, USA). From a set of seven U308 NBL reference standards available, NBL-98-1, 6 and 7 were not analysed as the expected amounts of the analytes in these standards were not covered by the present analytical range. The results obtained are shown in Table 4 along with the certified values. A comparison of these results indicates close agreement for all the determinations except for the determination of Ag in the 98-5 standard. The results therefore infer that the accuracy of the method is good.Conclusions Atomic absorption spectrometric methods using a graphite tube are described for the direct determination of Ag ( 0 . 1 4 p.p.m.), Be (0.1-2 p.p.m.), Ca (25-500 p.p.m.), Fe (2.5-100 p.p.m.), Pb (1-20 p.p.m.) and Sn (1-50 p.p.m.) in uranium solution. The 5 pl of sample used in these determinations contains 100 pg of uranium in a nitric acid medium. By matching the composition of the standards and the samples, a precision of 9% RSD is obtained in this work. Studies to investigate the effect of residual uranium in the graphite tube and that of the 18 concomitant impurities on the analyte absorbance showed the absence of any signficant interference. In view of higher concentrations permissible for Ca, Fe, Si, W and Zn in fuel samples, their effect on the analyte absorbance has been studied at concentrations of 1000 p.p.m. A measure of the high accuracy of the analyte determinations was obtained from the analysis of four U308 NBL reference samples. The authors are grateful to Dr. M. V. Ramaniah, Director, Radiological Group B.A.R.C. and Dr. P. R. Natarajan, Head, Radiochemistry Division B.A.R.C. for their interest and encouragement during the course of this work. References 1. 2. 3. 4. 5. 6. Baudin, G., Progr. Anal. At. Spectrosc., 1979,3,2. Buffereau, M., and Robichet, J., Methods Phys. Anal., 1971,7, 138. Patel, B. M., Bhatt, Paru, M., Gupta, N., Pawar, M. M., and Joshi, B. D., Anal. Chim. Acta, 1979, 104, 113. Patel, B. M., Gupta, N., Purohit, P., and Joshi, B. D., Anal. Chim. Acta, 1980, 118, 163. Patel, B. M., Goyal, N., Purohit, P., Dhobale, A. R., and Joshi, B. D., Fresenius 2. Anal. Chem., 1983, 315, 42. Recommendations by Commission on Spectrochemical and other Optical Procedures for Analysis on “Nomenclature, Symbols, Units and their Usage in Spectrochemical Analysis- 11,” Pure Appl. Chem., 1976, 45, 99. Paper J6l25 Received April 9th, 1986 Accepted March 6th, 1987

ISSN:0267-9477    

DOI:10.1039/JA9870200459

出版商:RSC    

年代:1987

数据来源: RSC